AMD's Ryzen CPUs (Ryzen/TR/Epyc) & Vega/Polaris/Navi GPUs

There is a boosting bug holding back the Ryzen 3 boost frequencies, the bug not seen in a previous BIOS with 3900x boosting to 4.65ghz. So perhaps the AMD PB2 video wasn't such a tease after all (not just for 3950x) and we haven't seen the full capability of Ryzen 3 as shown in the current reviews and benchmarks.

Windupmykilt 7 hours ago
@Gamers Nexus
..."It looks like there are a couple of AGESA bugs and possibly a driver bug. The 3900X is seen hitting 4.65 on AIOs with the original BIOS. Should be interesting to figure out the problem cause this most certainly is hampering benchmarks."

Rumor: PSA: Ryzen 3000 Gaming Performance is being gimped by MB bios issues. Explains inability to reach advertised boosts. xanxogaming.com
Submitted 15 hours ago by HerpDerpMcChirp
https://www.reddit.com/r/Amd/comments/cacwf9/psa_ryzen_3000_gaming_performance_is_being_gimped/
https://www.reddit.com/r/AMD_Stock/comments/caczu2/ryzen_3000_gaming_performance_gimpedboosts_not/

BaitForWenches 580 points 14 hours ago
"So people don't have to read through the whole thing, this is the part about the boost frequencies. (seems like benchmarks might need to be redone)"

Spoiler: The whole story…

During the first three hours of testing of the AMD Ryzen 9 3900X processor, using the X570 AORUS XTREME board, I noticed the problem when PCMark 8 did not pass the first test after 40 minutes (this is a total of ten tests). I noticed WHEA error (Windows Hardware Error Architecture) in HWInfo64 (se this software for PC telemetry, highly suggested).

From there I also decided to pay more attention to HWInfo64 and also checked that the BOOST frequencies of the processor had problems, since it didn’t get to “boost” all its cores to the maximum that it should, which is 4.6 GHz. It reached 4.5 GHz to 4.575 GHz in a pair of cores and the rest of cores to 4.3-4.4 GHz… We used manufacturers chipset driver, we have used press chipsets, as more current chipset driver version, same results.

It seemed strange to me, so I first decided to write to my contact with GIGABYTE USA ( Matthew Hurwitz, I thank him for all the time he has put in to find a solution) and showed him the WHEA (PCI Express) errors, as well as the rare behavior of the 3900X boost frequencies.

Midnight (Wednesday) GBT HQ gives us news and according to their tests, the new AGESA code, including NPRP BIOS (BIOS for press) replicated our results in single-core frequencies, BUT, the original BIOS (AGESA 1002,without code introduced NPRP) turbo boost was working well.

With this information, I decided to flash BIOS, the first BIOS released for the X570 AORUS MASTER board and surprise, the boost frequencies were working as they should, even beyond the processor at 4.65 GHz. The WHEA error problem in the PCI Express was still going on, so I kept pressing and trying if the problem was maybe the chipset driver.

AMD Ryzen 9 3900X Review & Benchmarks: Premiere, Blender, Gaming, & More
Gamers Nexus
Published on Jul 7, 2019
For the AMD R9 3900X CPU review, we're focusing first on production workloads (Adobe Premiere, Photoshop, Blender, V-Ray, et al.) and secondarily on gaming.
Watch our Ryzen 5 3600 review - we strongly recommend the 3600: https://www.youtube.com/watch?v=7AbNe...
Our RX 5700 XT review is here
Article is pending -- we're exhausted from putting together 5 reviews in a day, so give us some time on the articles. If you want the 3600 article in the meantime, it's here: https://www.gamersnexus.net/hwreviews...


07-07-2019, 06:01 AM - The Stilt
Strictly technical: Matisse
"First and foremost, a word of warning. When reading ANY of the AMD Ryzen 3000-series "Matisse" launch-day reviews, the first thing you should do is navigate to the page which lists the hardware setups.
AMD supplied four different motherboards to the media, one from ASRock, ASUS, GIGABYTE and MSI. In case of the ASUS Crosshair VIII Hero Wi-Fi motherboard, the media was instructed to use 0066 bios build, which had been vetted and approved by AMD. However, newer bios builds were available and ASUS has also (allegedly) told the media to use those versions. What exactly has transpired here is still under investigation, but regardless of the actual reasons behind it, the consequences might be rather significant. In practical terms, all reviews which were done on ASUS Crosshair VIII Formula or Hero motherboards using other than 0066 bios build must be considered invalid, at least partially. Reviews using other ASUS motherboard models (not provided by AMD) are under suspicion as well."
https://www.overclock.net/forum/10-amd-cpus/1728758-strictly-technical-matisse-not-really.html

AMD Ryzen 7 3700X + Ryzen 9 3900X Offer Incredible Linux Performance But With A Big Caveat
Written by Michael Larabel in Processors on 7 July 2019. Page 1 of 11. 77 Comments
https://www.phoronix.com/scan.php?page=article&item=ryzen-3700x-3900x-linux&num=1

AMD tested on Ubuntu LTS 18.04 and Ryzen 3 boots fine, other recent releases of other distro's aren't booting:

"To jump straight to the new stuff (Linux support and performance)... On the likes of Ubuntu 18.04 LTS, these new AMD Ryzen 7/9 series 7nm processors deliver incredible performance with the newest AMD X570 motherboards. The support on the likes of Ubuntu 18.04 LTS is great except for possible compatibility issues depending upon the particular motherboard/ASICs employed.

So what's the problem? On newer Linux distributions, there's a hard regression either within the kernel but more likely some cross-kernel/user-space interaction issue leaving newer Linux distributions unbootable."

Michael
Phoronix #5 07-07-2019, 09:30 AM
Originally posted by Nille_kungen View Post
" Very impressive but i still was expecting some game benchmarks."

"Game benchmarks are coming, ran short on time due to spending easily dozens of hours trying to figure out the Linux systemd/boot issue."

Ryzen 3000, Booting Linux, and You, Software & Operating Systems, Linux, ryzen-3000
https://forum.level1techs.com/t/ryzen-3000-booting-linux-and-you/144702/18

RYZEN 3000 & RADEON "NAVI" 5700XT -- READY FOR LINUX?
JULY 7, 2019
https://www.level1techs.com/video/ryzen-3000-radeon-navi-5700xt-ready-linux

AMD Radeon RX 5700 / RX 5700XT Linux Gaming Benchmarks
Written by Michael Larabel in Graphics Cards on 7 July 2019. Page 1 of 7. 39 Comments
https://www.phoronix.com/scan.php?page=article&item=radeon-5700-linuxgl&num=1

As always, Linux will take a while to sort itself out with new Video cards, CPU's, and motherboard chipsets. Be Patient.

Silicon Lottery Ryzen 3000 CPU's ETA July 19th...
https://siliconlottery.com/collections/all

• AMD R7 3700X
• AMD R7 3800X
• AMD R9 3900X

 

From that post..
A word regarding the "Auto Overclocking" feature...

The new "auto overclocking" feature, which is advertised with up to 200MHz frequency increase, in reality does close to nothing, at least on higher-end SKUs.
The lower-end SKUs, such as Ryzen 5 3600 definitely get some advantage however, the higher-end SKUs such as the 3700X and 3900X can be completely maxed out simply by increasing or removing the power limit (through PBO).
These SKUs are already clocked so high that further frequency improvements theoretically made possible by the "Auto OC" feature are disallowed by the silicon fitness monitoring feature (FIT), due to the required voltage for higher frequencies being too high. For instance,
on the 3700X test sample the best core of the CPU raises its frequency by 25MHz when the highest 200MHz option is selected. The rest of the seven cores remain at their default frequency, which varies between 4.35GHz and 4.375GHz.
Meanwhile the 3900X, which has stock max boost of 4.65GHz, there are no gains what so ever. In fact, none of the cores within this CPU even reach the advertised 4.65GHz. The two best cores reach 4.575GHz, while the ten other cores reach 4.325 - 4.4GHz peak. The variation between the different cores even on the same piece of a silicon appears to be huge, which would indicate that the process isn't very mature at this point. Even AMD themselves state in their slides that the frequencies are limited by the voltage they can safely feed to the CPU.

The overclocking capabilities

Essentially, if we're talking about the higher-end SKUs, there is basically none.
Based on my experience, the best case of scenario on 6C CCDs (3600, 3600X and 3900X) is around 4.25GHz, at relatively safe voltage levels.
In case of 3900X, given that you can cool the chip with two of those 6C CCDs. SKUs with 8C CCDs (3700X, 3800X and 3950X) the best case is around 4.15GHz. The 3950X is expected to be thermally limited, as a whole.
The biggest limit is the intensity (heat per area), secondly the voltage you can safely feed to the silicon. For example, the 9900K which has a reputation of being an inferno, has theoretical intensity of ~1.15W/mm² when operating at 5.0GHz (200W @ 174mm²).
Meanwhile Matisse can easily reach intensity of > 1.5W/mm² (120W+ @ 74mm²). The second issue is, that beyond ~3.8GHz the V/F curve becomes extremely steep. According to FIT, the safe voltage levels for the silicon are around 1.325V in high-current loads
and up to 1.47V in low-current loads (i.e ST), depending on the silicon characteristics. Because the stock boost operation is already limited by the silicon voltage reliability, the only way to eke out every last bit of all-core performance is using OC-Mode. Like on previous Ryzen generations, entering OC-Mode also means that you will loose the turbo boost (all cores operate at same frequency). On the higher-end SKUs, the single threaded performance penalty will be massive from doing so. For example on 3900X, you'd be trading additional ~100MHz all-core frequency to a loss of up to 450MHz in ST frequency by doing so. Personally, I advice against overclocking the higher-end SKUs at all, and instead increasing the power limits and trying your luck with the "Auto OC" feature (which most likely isn't beneficial).

The V/F testing was done using full resource utilization (FRU), meaning the stability was tested using 256-bit workloads.
Unlike Intel designs, Matisse does not feature an offset for 256-bit workloads. This means that to ensure the stability of the CPU cores in every scenario, they must be tested using this kind of a workload.
On Matisse, the delta in power consumption between the scalar and 256-bit vector instructions is massive, as expected (37%). That being said, there seems to be other design related factors limiting the maximum achievable frequency.
Despite significantly lower power consumption and therefore also lower temperatures, stability even in pure scalar workloads could not be achieved at much higher frequencies, compare to FRU scenario.

This probably explains nicely how Hardware Unboxed fried a 3900x in a day. A lot of reports stating 1.5v being applied by default with the BIOS. 1.5v is just stupid for any extended period of time.

 

https://twitter.com/andreif7/status/1148170909322293248?s=09

 

Yes, there are the Asus auto CPU voltage overvolt bugs, separate from the AMD boost bug(s).

That's why I was pointing out that the boost results aren't correct as currently shown, making the review results skewing toward under reporting Ryzen 3 performance. After the bugs are resolved the Ryzen 3 performance should go up - IDK how significant the difference will be, but for now the great Ryzen 3 results being shown should improve further.

BTW, I've experienced these Asus BIOS auto CPU voltage overvolting problems before - warnings go out in the OC tuning community each time - over many years and many times - you can't use the auto setting for voltage in Asus motherboards - I disuse them in any make motherboard as they always overvolt to some extent, so this isn't new or unique to Ryzen / x570 - it's an Asus "traditional" issue that also to some extent affects all motherboard BIOs.

Andrei F.‏ @andreif7
"AMD Update: MSI yesterday has released a new firmware update for the MEG X570 ACE MB which we used for our Ryzen 3000 review article. https://www.anandtech.com/show/14605/the-and-ryzen-3700x-3900x-review-raising-the-bar/5 … We've noticed changes in the boosting behaviour of the chip; @gavbon86 is currently re-testing our whole test suite."

Yup, and it might take a while to work out complete resolution as they iterate testing on new updates from each board maker. Hopefully it will get resolved quickly - reports of earlier firmware not showing this boosting bug are encouraging.

Many of the reviewers were rushed to complete the reviews in time for release, including Anandtech, and now there will be plenty of time to take the time to redo the reviews and comparisons with updated (fixed) firmware, drivers, patched (fixed) OS's. As expected there is a long tail to new releases to get bugs fixed for the best and most accurate results from testing.

 

PCWorld coverage is really long this time with the insights from their testing over the last week of the AMD Ryzen 3000 CPU's 8:34 and AMD Navi GPU's 55:30

Reviews galore: Ryzen 9 3900X, Radeon RX 5700/XT | The Full Nerd ep. 99
Started streaming 12 minutes ago, 420 watching now
PCWorld
Join The Full Nerd gang as they talk about the latest PC hardware topics. Today's show covers all the major reviews including AMD Ryzen 9 3900X, Radeon RX 5700/5700XT, and Nvidia Geforce RTX 2060 Super and 2070 Super. As always we will be answering your live questions so speak up in the chat.
Ryzen 3000: 8:34
Radeon 5700: 55:30


The Navi 5700 performance results looks great as compared to the Vega 56 / 64 - outperforming both, with exceptions - like Metro Exodus - it's clear some AMD driver updates to address specific game performance shortfalls are needed:

RX 5700 vs RX VEGA 56 Benchmarks | Gaming Tests Review & Comparison | 53 tests
Benchmark PC Tech
Premiered 18 minutes ago

Spoiler: Tests performed:

AMD RX VEGA 56 vs AMD RX 5700 BENCHMARK REVIEW / DX12 INCLUDED / GAMING TESTS – RTX 1650p / 1440p / 4K Games tested: BATTLEFIELD 5, DEUS EX: MANKIND DIVIDED, FAR CRY PRIMAL, GEAR OF WAR, GTA 5,HITMAN 2,SHADOW OF THE TOMB RAIDER, ASSASSIN'S CREED ODYSSEY, TOTAL WAR: WARHAMMER, WATCH DOGS 2, WITCHER 3, STAR WARS BATTLEFRON 2, DESTINY 2, HELLBLADE SENUAS SACRIFICE, MIDDLE EARTH SHADOW OF WAR, PREY, PROJECT CARS 2, FAR CRY NEW DAWN, METRO EXODUS, 3DMARK FIRESTRIKE, 3DMARK TIME SPY.

RX 5700 vs RX VEGA 64 Benchmarks | Gaming Tests Review & Comparison | 53 tests
Benchmark PC Tech

The 5700 doesn't quite make it to the performance levels of the Radeon VII, the 5700XT should be much closer to the Radeon VII in performance for much less cost - the Radeon VII could get a price cut.

RX 5700 vs RADEON VII Benchmarks | Gaming Tests Review & Comparison | 53 tests
Benchmark PC Tech

The 5700XT Benchmark PC Tech tests are coming soon, I'll add them here when they post.

RX 5700 XT vs RX VEGA 56 Benchmarks | Gaming Tests Review & Comparison | 53 tests
Benchmark PC Tech

RX 5700 XT vs RX VEGA 64 Benchmarks | Gaming Tests Review & Comparison | 53 tests
Benchmark PC Tech

RX 5700 XT vs RADEON VII Benchmarks | Gaming Tests Review & Comparison | 53 tests
Benchmark PC Tech

AMD Radeon RX 5700 XT & RX 5700 Review, New Value Champions!
Hardware Unboxed
Published on Jul 7, 2019

 

https://www.reddit.com/r/Amd/comments/cajvsd/b450x470_owners_who_picked_up_ryzen_3x00/

 

The M/B mess AMD has created for itself is sad and needn't have happened at all (AMD, just release a (single) new M/B platform already). The other glitches are par for the course for me whenever I've tried testing any AMD platforms previously. This is far from an AMD dominated landscape for anyone that needs things to work.

Again, glad AMD delivered so much of what it boasted/marketed about. But ultimately, this generation is a hard pass (so far).

See:
https://www.tomshardware.com/news/amd-ryzen-threadripper-2-cpu-sale-deal,39810.html

The deals are available though on both sides of the fence.

 

I do not consider it a mess as you describe. This has been an issue with every new generation, the bios flashback issues. Meaning this really needs to be done by tech experts, something though AMD should explain from the get go.

Considering the size of the launch I think the problems are few and far between. and for the most part are upgrading existing systems. Anyone replacing to a new system probably will not have all these issues.

 

all early adopter stuff people have to deal with. intel has been on the same stuff for decades and MB people are use to it, socket/chipset change with almost nothing changes etc. imo it is understandable issue will arise with new stuff, intel will have that too when its new arch comes too.

i'll simply wait and wait, for zen 3 or 4, or tigerlake 10nm++

 

Der8aur on the X570 with Gen 3 SSD and Gen 4 SSD HW (Gen 4 temperature claims.), and It is to be really noted as how the X299 also doesn't have any Active cooling. Also if you guys looked at that abomination of ugly MSI X570 Godlike (~800Eur) with all that stupid plastic around the repurposed Matisse I/O chipset PCH (plus only 6 SATA III Storage option). Almost all companies did this, except one SKU, Aorus Extreme X570 with massive proper Metal HS without all that tacky rubbish plastic, given how they mentioned the PCB layers and traces for the Gen4 signals effecting the cost but couldn't fit a proper HS and instead block it with cheap plastic, shame.

Edit : Unfortunate that there is no space left for OCing or tweaking Zen 2, perhaps UV like how GN did. AMD extracted every single drop out of these chips pre-binned OOB, Auto OC and PBO and set. It's cool for a normal consumer but for a novice or new comer tinkerer like myself it just kills all the fun out, Intel 9900KS even if its 5GHz OOB there is still room left. With the Der8aur delidding video he shows how the Ryzen 3000 scales with the low temperatures, also the voltage..Oh and the DDR4 4000+ is just a damn sticker, useless since it hampers IF speeds, 3600 is max again.

9900K/9700K price reduction would be good IMO, plus the 10C/20T on CFL might not happen due to the heat ? and my only worries are how the 2020 14nm++ Cometlake/Rocketlake would position in the Z370/390 compatibility area if we go with Z390, Intel needs to be careful tbh with this 10nm is totally out of equation and 1.5 years without a mainstream desktop CPU will be hemorrhaging, so 14nm++ is for sure on the table given it's ultra optimzations.

Going with X570 with 3700X or 3800X (Absolute zero OC I guess with this) would be great. Since Zen 3 might catch up with the remaining gap on 7nm+ and more uarch improvements (?) so an upgrade to the top end SKU in 2020 would be great rather than buying that expensive 16C 3950X part today.

I don't think there will be any refresh to the AM4 chipset again since 2020 is the purported max support, again refresh means kick in the gut for these expensive X570 Mobo buyers, like Nvidia Super (?) 2021 might see DDR5 or Gen5, both I don't expect it to happen.

Edit 2: I feel AMD took a page out of that Intel Broadwell i7 5775C eDRAM and gave the CPUs that huge L3 cache for gaming workloads along with the overall processor perf . Plus I wish Intel to ditch that iGPU entirely and give the mainstream K chips some benefit with precious space like more cache with their solid Ringbus which has solid advantage over their Mesh or the AMD's MCM IF for 2020 lineup, there's a big but again, XCC, HCC, LCC and Xeon blah blah manufacturing expenditure since they already announced Cascade Lake-SP and AP..sad

 

Sorry but I prefer the throw it in and forget it method with max performance out of the box. That is as long as the performance is killer from the get go.

 

PSA: 5700 Series Custom AIB Designs
Submitted 5 hours ago by scottherkelman VP & GM Radeon Business Unit
https://www.reddit.com/r/Amd/comments/catck3/psa_5700_series_custom_aib_designs/

"Hey all, Custom AIB designs will be hitting the market ~mid August"

And, we have the "pricing" 4D Chess thinking explained and confirmed:
https://www.reddit.com/r/Amd/comments/c9sm4n/excellent_explanation_about_the_navi_price_cut/

DarkKratoz 53 points 2 days ago
"AMD gets close to launching Navi. Knowing Nvidia's history, they can guess that Navi will be immediately countered by Nvidia (like when Vega met the 1070Ti, or when the 590 met the 1060GDDR5X refresh). Super is teased. Navi is shown off, given some performance numbers and a price that everyone feels is inflated. It's low hanging fruit for Nvidia, and if purposefully done, draws out the Nvidia Super cards at a certain price point as an answer to Navi. If Navi was always intended to be priced around the new, current prices, then AMD purposefully baited Nvidia, because Nvidia rarely drops the price of their hardware for the first half year. Editorializing; Nvidia always tries to steal the price/performance thunder of AMD whenever AMD is ready to launch a great new card. It's not hard to believe that Lisa, wanting to get Navi into the public's conscience the same way that Zen 1 was, wanted to get ahead of Nvidia's counterattack without adjusting their plans from the start."

scottherkelman VP & GM Radeon Business Unit 18 points 2 days ago
"You nailed it, congrats!"

Asus provided a BIOS fix that works for their motherboard:

07/08/2019 6:33 PM (GMT) - Update on the bios issue on Crosshair VIII Hero motherboard ("the thing").
https://www.overclock.net/forum/10-amd-cpus/1728758-strictly-technical-matisse-not-really.html

The current state of PBO on Ryzen 3000
Submitted 5 hours ago by anethma
https://www.reddit.com/r/Amd/comments/casxjs/the_current_state_of_pbo_on_ryzen_3000/

Ryzen Boost Clocks vs. BIOS: AMD AGESA 1002 vs. 1003a/b Differences
Gamers Nexus
Published on Jul 8, 2019
A lot of people have been asking about how AMD's boosting behavior performs in the Ryzen 5 3600 and Ryzen 9 3900X. There are no differences at all in our 3600 results and 3900X is barely changed. This benchmark "retests" the AMD R5 3600 and R9 3900X CPUs with three different BIOS versions on the Gigabyte X570 Master motherboard. This will vary CPU-to-CPU and board-to-board, but we saw 0 difference on the 3600 and very little on the 3900X. Note that all-core is 100% unaffected in our tests and remains identical. The reviews are unaffected and conclusions remain the same. Other reviewers may have a bigger impact. It'll depend on the board, BIOS, and the CPUs used.

For those of you struggling with your Ryzen 3 on x4xx:

Unable to Boot New Desktop System Configured with AMD 3rd Generation Ryzen™ Desktop Processor, and AMD Socket AM4 400-Series Motherboard
https://www.amd.com/en/support/kb/faq/pa-100

BIOS Update For New Ryzen 3000 Processors – x470 & B450
https://www.wepc.com/tips/bios-update-for-new-ryzen-3000-chips-x470-b450/

AMD's Scott Herkelman Appreciation Thread
Submitted 10 days ago by 808hunna
https://www.reddit.com/r/Amd/comments/c6klxj/amds_scott_herkelman_appreciation_thread/

scottherkelman VP & GM Radeon Business Unit 22 points 10 days ago
Hey thanks for this, I swear we have the best community.
We can’t always talk about everything, but we can definitely improve how we communicate. That’s why I visit this site everyday. I can’t promise to respond but will be reading a lot.
So please keep reminding us when we are not clear or may have missed something - compliments are nice too

Slacker Live" Stream 004-Overclocking the new AMD Ryzen 3700x
bearded hardware
Streamed live 9 hours ago
Almost 6 hours of Memory OC'ing...

AMD Zen 2 Memory Performance Scaling with Ryzen 9 3900X
https://www.techpowerup.com/review/amd-zen-2-memory-performance-scaling-benchmark/

This is The Best Memory For Gaming on Ryzen 3000!
Boot Sequence
Published on Jul 8, 2019

Spoiler: Build advice...

Ryzen 3000 is finally out, And you might want to build a brand new rig, or even update the one you have right now.well, theres something that a lot of people overlook when buying computer parts and I have to admit, I’ve overlooked it myself. That thing is memory and today I’m diving deep into the perfect memory for gaming on Ryzen 3000! What is up guys i’m snows and this, is .. a different kind of boot sequence. So you chose your CPU, it could be the R5 3600 which seem to be the perfect chip for gaming, or it could be the 3900x if you wanted a lot more cores and power. You chose your case, your motherboard, your power supply, your storage,your RGB but now its time to make a choice about memory. Do you go for the cheapest 2133Mhz kit you can find, or do you go all out and buy a 4000+Mhz kit to make sure its the fastest it can be. Well, while ryzen 3000 has no problems running super fast memory, it doesn’t mean it's the best choice Or the best bang for your buck. For the first 2 generation of Ryzen , the memory clock was synced one to one with the infinity fabric. So if your memory was at DDR4-2133 which is actually 1066Mhz,remember DDR means double data rate, your infinity fabric would also be at 1066Mhz. So the higher your ram speed the higher the infinity fabric. But it started to get unstable the higher you got which meant that DDR4-4000 on ryzen was extremely hard to achieve. Now, with Zen 2, the memory can be decoupled very easily at a certain point. That specific point is between 3733Mhz and 3866. Once you cross that line you Infinity fabric speed get’s halved to compensate. At that point things like raw memory latency take a hit. But does it really matter ? Does latency truly make a difference in games. Well yes it does and Tech power up went through rigorous testing to find out which memory speed is best for Zen 2.

 

i would be interested to see how intel will target this issue with their 10nm++. based off their yrs old slides that 10nm++ will perform better than 14nm++ in terms of efficiency which is very much that V/F curve wall. with TSMC, AMD will win in server market with it's chiplet design for now but consumer space will have a hard time to top intel.

intel's 8-12 cores monolithic die for OVERALL (including ram sadly) latency, high frequency, AMD is pretty much stuck at around 4.3-4.4ghz at best.

 

Nvidia dirty tricks...? Proof you can't disclose...? Do tell...

AMD's Navi is Keeping Jenson Up at Night, so Nvidia is fighting Dirty!
Moore's Law Is Dead
Premiered 9 hours ago
Did you notice some channels came to odd conclusions with Navi? There is a reason this is happening again... & there is a reason Nvidia is so scared of Navi.

Spoiler: TIMESTAMPS:

1) Navi IS impressive if you look at the NUMBERS 0:27
2) Why some channels aren't shilling 2:42
3) Nvidia is up to old tricks 3:42
4) Why RDNA really is Impressive as an Architecture 4:31
5) What's different this time that terrifies Nvidia 8:20
6) A direct Message to Other Tech Channels 11:17
7) One more thing about my Channel 13:29

Jason McGrody 8 hours ago (edited)
"My basic conclusion is Ryzen 3000 is outstanding and Navi is good (largely because of prices that Intel and Nvidia choose for their products and a mediocre blower cooler). BUT if you have a Microcenter near you then you can get $100 of Ryzen 3000, motherboard, and Navi card. Ryzen is so good it's worth buying at MSRP. So, basically, (at least it's the way I think of it) you get Navi for $100 off. Navi 5700XT for $300? Are you kidding? Are. You. Kidding?"

Joe Richard 9 hours ago
"I was going to buy the GTX 1080ti used and now that I saw those benchmarks, I believe I need to re-think my thoughts. Thanks "

Daniel Soutar 9 hours ago
"I feel sorry for those channels you describe. But I feel even more sorry that a video like this even needs to be made. Keep up the great work, and kudos for calling out poor form on Nvidia's part in bad-mouthing the competition."

I was wondering why I was hearing so much about Nvidia, Ray-tracing, and DLSS inside of Navi reviews...

RX 5700/XT Getting a RAW Deal From the Media... on Purpose?
The Good Old Gamer
Published on Jul 8, 2019
After the craziness that was the Ryzen 3000 and Radeon RX 5700/XT launch I've had time to examine more reviews out there. Some of what I found was simply SHOCKING! Add to that I learned some "Reviewers" were approached by Nvidia to influence their NAVI reviews.


Apparently R9 290 / 390 Coolers are a match fit for the 5700 series.


https://twitter.com/HardwareUnboxed/status/1148113823137533952

https://twitter.com/watchtimmyjoe/status/1148026906991833088

https://twitter.com/watchtimmyjoe/status/1148367829969920001
https://www.reddit.com/r/Amd/comments/cakxlq/hardware_unboxed_on_twitter_i_put_an_r9_290/

 

At least Nvidia knew how to make reference blower design that didn't blow suck.

 

I agree, why waste time tweaking for hours for little real world usage gain. Daily driver settings for reliable performance and stability is much more interesting.

AMD automatic tuning does a better job for single core but for multi-threaded work you can get a slight boost overall for a small loss in single threaded tasks.

Jarrod shows the difference between stock and overclocked for all of his Ryzen 3 comparison tests, here's one example:


AMD Ryzen 7 3700X vs Intel i7-9700K - Which 8 Core CPU In 2019?
Jarrod'sTech
Published on Jul 8, 2019
How does AMD’s new Ryzen 7 3700X CPU compare against Intel’s i7-9700K? I’ve benchmarked games at 1080p and 1440p resolutions to show you the gaming differences, both at stock settings and while overclocked. I’ve also tested a number of other applications including rendering and video editing software suites like Adobe Premiere. We’ll also compare temperatures and power draw to help you decide which CPU is worth it for you.


PirateKing Strawhat 20 hours ago
"I'm watching these charts and it looks like there is no need to over clock the 3700x, I'm loving this even more now"

Jarrod'sTech 20 hours ago
"I'd just stick PBO on for games and relax "

Jarrod also put together a 3900x vs 9900k comparison with similar stock vs OC results:


AMD Ryzen 9 3900X vs Intel i9-9900K - CPU Comparison
Jarrod'sTech
Published on Jul 7, 2019
How does AMD’s new Ryzen 9 3900X CPU compare against Intel’s i9-9900K? I’ve benchmarked games at 1080p and 1440p resolutions to show you the gaming differences, both at stock settings and while overclocked. I’ve also tested a number of other applications including rendering and video editing software suites like Adobe Premiere. We’ll also compare temperatures and power draw to help you decide which CPU is worth it for you.

 

Hopefully with with process maturity and the 3950x higher sustained clocks will be achievable. Maybe once we start seeing the TR3's?

Also the 3900x beating a 7980xe, can't wait for the 3950x and tr3's.

 

Therein lies the problem. Clock for Clock, the new AMD chips => Intel chips IPC wise, but their clock rates tap out way below Intel (as De8auer showed), so even a mild OC puts Intel back in front core for core. Not earth shatteringly so, but it is still there.

Overclockers are outliers, so the the vast majority of users, the 3900x > 9900k.

I'm all for "MOAR COREZ PLZ!1!1!", but for the bulk of day to day use, they lose their appeal after a certain core count depending on use cases and it comes back to IPC and core for core performance. All Intel is going to do now is continue to release clocked up 9 series chips and they still have the performance lead core for core.
If the 3900x had the ability to clock to even 4.8 on all cores, I would be all over it. I'm definitely keeping an eye out on the 3950x and Intel should definitely be worried. I remember when the FX-60 was beating Intel back in the day and switched to AMD in a blink of an eye. Get your OC core for core performance up, and I'll gladly do it again.

Intel has been languishing for awhile, failing to deliver on 10nm and failing badly in the modem space.

As for the 5700/xt, another solid winner in the middle tier for AMD that has and will continue to force Nvidia to adjust pricing and offer mid tier bumps in performance. In other words, business as usual. I love my Vega 64. It was dirt cheap brand new when I got it (349.99), and it meets my performance criteria. I saw the prices of the 2080/2080ti and ran the other way. I might switch to a 5700xt based on performance comparisons to the 64. I haven't seen a head to head yet.

 

dono about that looks like AMD going full course ahead to me. 7nm+ is prob the last gen we'll see some acceptable clocks unless TSMC somehow pulls out a miracle on the 5nm node by default.

and who would have know that AMD's own chipset made by themselves can compete against intel in storage performance. @ajc9988 the performance jump from their x400 series mobo chipset to 500 series chipset for storage was a big one, comparable to intel in storage now.

 

I posted some a couple of pages back, and here in the middle of this post there are 5700 / 5700XT vs 56 / 64 / VII, and that channel has a lot of other comparisons:
http://forum.notebookreview.com/thr...ga-polaris-gpus.799348/page-631#post-10930327

There are only so many that can afford the 9900K or 3900x, there may be far more than purchase the 3700x / 3800x which does well against the stock 9900k and even the 3600 / 3600x are no slouches for far less $.

If the Ryzen 3 OC'd / boosted to past 9900k performace and exceeded the 9900k in gaming definitively, it might wake the sleeping giant. Intel might even get desperate enough to contract out for 7nm desktop, laptop, server CPU's, and then AMD would have to pull out all the stops. It would be chaos.

Letting Intel keep the Crown in a bit of a shared custody agreement - AMD costs less - Intel gets a 5% token win in games - then AMD can continue to chomp away at Intels marketshare.

It's a profit thing, over the long haul.

That is unless Intel fails to deliver 10nm again, and again again, then it might not matter and AMD can just keep iterating past Intel without worries.

I hope you enjoy the 5700 / 5700XT, check back a few posts - it looks like the R9 290 / 390 coolers fit the Navi GPU's, as well as AIO and custom block coolers, and that might be a great project / long term GPU.

As told by Tiny Tim Logan from OC3D the Aorus PCIE 4.0 SSD performed at full performance through the x470 MB - same speeds as seen on the x570 model... click the link and listen...starting at 34:00

 

A very thorough and fair review of the 5700/5700xt really showing it is a damn good card, outclasses the Vega56 /64 and is on average @1440p (I run @ 2560x1600):

~9% slower than the 2080
~6% slower than a Radeon VII
~2% slower than the 2070 Super

@ a $400 price point.

I went ahead and ordered one because I like what I see and I'll run some personal tests to see how it stacks up in my config. I always thought the VII was too
expensive @ $700. The 2080ti is way out of my modern day budget on all fronts.

https://www.techspot.com/review/1870-amd-radeon-rx-5700/

 

i need to see it for myself screenshot to believe it as most people says full peformance, as in sequential read/write of the ssd being advertised. what im after is 4k random read/write, and other metrics, vs the same SSD on the intel chipset.

no one really test SSD on system anymore extensively for intel vs AMD comparison after Jon left tweaktown. its not a hard thing to do but requires one to have both intel and AMD systems, or at least an old zen+ and zen AMD system to test jump in performance of different chipset (by asmedia vs AMD).

heres a review that Jon did https://www.tweaktown.com/articles/8073/amd-ryzen-ssd-storage-performance-preview/index3.html

 

He mentions the PCIE 4/3 throughput numbers elsewhere in his videos, but I wish he would have published the benchmark runs with the full transfer results - like you want - to back up the claim.

There was a clear stepping back by AMD and board makers on support for PCIE 4.0 on x470, it's possible these numbers might not persist as BIOS's are updated and somehow block this from working, hopefully that doesn't happen.

I think enough people will be reporting in with their experiences soon enough, and some will publish their benchmark runs with PCIE 4.0 SSD's on x470.

Cool, and "Intel HQ" is likely a short distance from Microcenter's across America, and if you buy a MB / CPU / Navi you get $100 off of your purchase, making the 5700XT only $300!

 

LOL, I like the reference to my Intel HQ.

There is actually a Microcenter ~70 min from me in PA.

 

https://www.techpowerup.com/257201/...ail-amds-zen2-backwards-compatibility-promise

 

The Beard seems to be stuck without luck with the current MSI board "Freezing" when the temps get low on LN2, a couple of hours in with no joy. Could be BIOS, could be pot problems, he's going back and redoing it all over again with a new BIOS. No success, a new BIOS may be needed.

"Slacker Live" Stream 005-Overclocking the new AMD Ryzen 3700x on Liquid Nitrogen (LN2)
bearded hardware
Started streaming 2 hours ago

 

RVII is heavy on the compute side of things (comparable to 2080ti and some Titans - plus the FP64 on this GPU is enormous, biggest of any other consumer grade GPU on the market - thanks to being a re-purposed data centre GPU) and comes with 16GB HBM2 while being a capable gaming GPU.
Strictly speaking, its not just a gaming GPU.
Taking everything into account, its comparable in price to 2080 while being far more potent at productivity workloads.

HBM alone would jack up the price by around $200 vs say GDDR5 or GDDR6, but then again, don't think Vega was specifically designed to work with GDDR in the first place (or if it could).

Anyway, kudos on getting 5700XT.
Looking forward to your reviews.

 

I saw an article somewhere that said the memory alone accounted for nearly half the cost of the R7.

I hope AIB 5700/XT cards will be released soon as I'm not a fan of blower style coolers.

 

The Beard finally gets some results, and under the "Spoiler Timestamps" is a long detailed index into what occurred during his long LN2 session:

Record set:
2:26:17 5.35GHz at 1.85 - [CB15] 2835 VALIDATED
"...we found that AMD Ryzen Memory performance is key at the proper FCLK and MEM Dividers, basically 1900 on fclk and 3800 mem is the optimized point and getting the lowest latencies possible is key."

"Slacker Live" Stream 006-Overclocking-Ryzen 3700x at 5.2ghz+ 1.9+Volts
bearded hardware
Streamed live 10 hours ago
The normal "Slacker Live" stream will be every Thursday 4pm EST(NYC).

Spoiler: YT Details...

...Now that the bios issues are resolved and basic scores acheived, its time to push the limits. 1.6 was needed to 5.2, so lets test to 1.9 After messing with memory on the last 004 stream, we found that AMD Ryzen Memory performance is key at the proper FCLK and MEM Dividers, basically 1900 on fclk and 3800 mem is the optimized point and getting the lowest latencies possible is key. The rumors with fclk is that they cause issues with Cold bug on LN2, we will find out and see if we can break the 8 core records.

Spoiler: Timestamps

Pinned by bearded hardware
Krakup Kiwi 2 hours ago (edited)
Some timestamps i have so far
1:03:40 2793 Cinebench
1:07:28 2812 CRASHED
1:22:08 2804 in Cinebench R15 (Crashed)
1:37:00 FULL POT 5.3GHz 2974
1:39:40 RUN IT AGAIN
1:41:40 2801
1:42:55 5.35GHz at 1.8v - 2805
1:45:18 5.375GHz at 1.85v - CRASH
1:47:57 EMPTY POT AND RESTART
2:00:50 5.0ghZ AT 1.8v
2:24:32 5.3GHz at 1.8v and 1.85v - 2747
2:25:46 5.325GHz at 1.85v - 2792
2:26:17 5.35GHz at 1.85 - 2835 VALIDATED
2:28:45 5.375GHz AT 1.85v - CRASH AND REBOOT
2:30:28 5.375GHz AT 1.9v - CRASH AND REBOOT
2:32:26 5GHz AT 1.8v -2626
2:33:24 5.325 at 1.825v 2794
2:34:42 5.35GHz at 1.825 2792
2:37:00 IS IT DONE?? Time to heat up and raise the Infinity Fabric

Jay sits down and runs through OC'ing the 3900x, including dealing with Asus overvolting - correcting it before OC'ing, and he got some good results:

AMD 3900X Overclocking explained... How to get improvements!
JayzTwoCents
Published on Jul 10, 2019
Overclocking AMDs new Zen 2 based CPUs can be a challenge... here are our results from our days of testing multiple configs and what gave us the greatest improvements.


Finally, some 3800x benchmarks, but he didn't have his chosen ideal cooling solution ready, so even though the 3800x hit 4.5ghz in boost the sustained max was 4.2ghz all core limited by cooling, look for updates when he gets the cooling sorted:

My 3800X benches
Submitted 18 hours ago * by HylianSavior
https://www.reddit.com/r/Amd/comments/cbcjh7/my_3800x_benches/

HylianSavior S] 3 points 3 hours ago
"Hah, yeah. I arrived home with my 3800X and realized I didn’t have my AIO AM4 bracket and I was fresh out of Kryonaut, so I had to settle for 5 year old AS5. Ordered the bracket and some more paste, so I’m hoping for some improved PBO results soon!"

HylianSavior S] 14 points 18 hours ago
"Got it from Micro Center! They just came in stock this morning."

Ryzen 3 CPU responds to load with accompanying voltage increases due to monitoring tools interrogation @ 5x / second causing frequency boosts and voltage increases - keeping the CPU from idling most efficiently. Robert Hallock of AMD responded a bunch of times in this thread - lots of good reading. The summary includes a link to another thread with more info too:

The final word on idle voltages for 3rd Gen Ryzen
Submitted 3 hours ago * by AMD_Robert Technical Marketing
https://www.reddit.com/r/Amd/comments/cbls9g/the_final_word_on_idle_voltages_for_3rd_gen_ryzen/

"tl;dr: Observer effect bad. You can't always trust your tools. CPU-Z gives you the right idle voltage. We'll look at the rest. Thank you everyone for your reports and insight, which helped us get to the bottom of this once and for all."

AMD_Robert Technical Marketing 36 points 19 hours ago
"Your understanding is correct. The cores are on and boosted at the voltages you're seeing, or asleep and unreadable with normal tools.

//edit: One other point I forgot to make, is that Ryzen adjusts clocks and voltages every 1ms. Most tools only capture every 1000 to 2000ms, so they're not showing you very much of what Ryzen is doing. HWINFO can adjust the timing resolution down to 100ms if you wanted but, even then, it's only showing you 1 of 100 possible samples. And it's still not going to show you if the core is asleep or not."

AMD_Robert Technical Marketing 43 points 19 hours ago
"Worry not! Everything looks a-ok. I guess the key takeaway for anyone who stumbles by this thread in the future is that the CPU's own logic determines what voltages are appropriate for the current condition(s).

Those conditions include: temperature, time under boost, number of active cores, number of sleeping cores, the "weight" of the thread, VRM temperatures, and more. Lots of telemetry goes into the decision to pick a clockspeed and voltage.

We built that firmware. We don't let anything interfere with it unless you manually OC the chip. That firmware would never do anything risky with voltage, and we know that because we built it that way. "

Robert Hallock ‏Verified account @Thracks 5h5 hours ago
Seeing chatter about idle voltages on 3rd Gen Ryzen. Here's the skinny: Some of the monitoring tools have aggressive polling intervals. They frequently wake every core to ask for status, which the CPU interprets as "PLS BOOST."
https://twitter.com/Thracks/status/1149041978107211776

"DID YOU KNOW? The new version of AMD Ryzen Master can show you core behavior that other tools can't or don't! Examples include: cc6 sleeping cores, sub-2200MHz idle cores, sub-1V idle voltages, fastest CPU physical core, your motherboard's VRM capacity, and more."
https://twitter.com/Thracks/status/1148810420884561920

Here's compatibility tests with a B350 + Ryzen 3's CPU from Hardware Unboxed:

Ryzen 9 3900X On a B350 Motherboard, Does It Cook?
Hardware Unboxed
Published on Jul 10, 2019


Gamers Nexus's inexpensive DIY solution to assist GPU / heatplate pressure to improve cooling for the 5700 / 5700XT that doesn't require a new cooler, improving the existing cooler with new paste + more pressure on the cooler to the GPU:

Fixing the RX 5700 XT Cooler for $0.04 | Paste & Washers vs. Thermal Pad
Gamers Nexus
Published on Jul 10, 2019
We added some washers and thermal paste to the RX 5700 XT cooler to fix some of its previously dismal performance. We also used 4 washers, found easily. The idea is to fix the mounting pressure issue on the RX 5700 XT cooler, a similar issue exhibited by the Radeon VII cooler. This addresses some of the thermal and temperature/noise concerns by bringing down overall GPU temperature, thus allowing lower noise levels at the same performance.

 

Now that AMD is showing such great performance offerings they need some OEM wins. Once available in stock systems they can truly start to see some major market share.

 

HP and Dell will make then favours using DDR4-2133Mhz in single channel, using cooling solutions fit for 10Watts using 25Watts CPU's, like they usually do, so that the AMD equiped laptop will be sure to be slower and run hotter/louder than its Intel equivalent..

 

Maybe not this time. This time Intel doesn't have an equal performance / cost solution for vendors to fall back on - to focus on - and those same vendors have been messed around with by Intel on 10nm, not delivering when promised - causing vendors to waste time and resources on projects that never completed.

Self preservation should kick in allowing vendors to deliver functional and desirable AMD models with sane configurations and pricing.

IDK how soon this can happen, but I'm sure there are top engineers working on configuring new models, right now.

A 3900x/5700XT laptop would be awesome, huge, powerful, and not expensive?

Maybe a nice 3700x + 5700 would suffice, a nice balance of performance and cool operation.

MSI?, Asus?, AMD?

 

AMD RIS, or AMD Radeon Image Sharpening looks good and only uses 1%-2% of the FPS, and it works on [DX9, Vulkan, DX12 - no DX11 support, yet?] existing games, HU does a good job of running through the technology, testing games, and showing the improvement.

Radeon Image Sharpening Tested, Navi's Secret Weapon For Combating Nvidia
Hardware Unboxed
Published on Jul 11, 2019


RAY M 3 hours ago
"unlike NVIDIA Dlss game developer doesnt require to implement just go settings and turn it on and works. well you have my money AMD"

bennyg 2 hours ago
"So Navi sharpening has better game support, image quality, and performance on release than DLSS does after nearly a year in the wild. You could almost say of AMD's implementation... It just works?"

glostermeteor 2 hours ago
"Holy damn, such a simple solution kills nvidias DLSS which took soo many tensor cores and machine learning, bumping the price noticeably... Bravo AMD"

Radeon Image Sharpening Tested, Navi's Secret Weapon For Combating Nvidia
Submitted 7 hours ago by RandomCheeseCake
https://www.reddit.com/r/Amd/comments/cbvptp/radeon_image_sharpening_tested_navis_secret/

Aken_Bosch [score hidden] a minute ago
"For those who will play on next gen Zen+Navi APU this thing is a godsend. You will be able to take 900p medium and upscale it to 1080p + sharpening."

 

I think that might have been true a couple of years ago, even maybe as recently as last year, but most OEMs seem to be embracing Ryzen recently. The Thinkpad E495/595 are pretty much the same as their Intel counterparts, as is the Envy x360. Sure, many OEMs still use slower and/or single-channel RAM in the default configurations but even with those things it is not like Ryzen performs significantly worse than comparable Intel products. It doesn't perform to its full potential but it isn't crippled to the point where the average user is going to feel like Ryzen is worse than Intel processors.

 

TimmyJoe put together a Navi + R9 390 cooler and it is an interesting result, with lower noise (Navi noise wasn't bad in his system, see video), much better cooling at 100% fans with a nice bump in Firestrike score, start at 09:25 to jump the the fun stuff:

Visiontek Radeon RX 5700 2 Fan Cooler TESTED!
Timmy Joe PC Tech
Published on Jul 9, 2019
I made this partner AIB Rx 5700 card, ok I know but I think this was an interesting experiment to see how even a mediocre partner card will perform!


The other take away from his experiment is although better cooling does improve the scores, there is a boost cap that limits performance - he speculates the real AIB cards - some at least - will provide better cooling and overclocking headroom, but that's yet to be seen.

So if you are planning on putting a water block on the AMD Navi's, maybe wait for someone else to do it and see if they figure out how to break through the OC limits, wait for AIB cards, or enjoy the cool running better performance of a waterblock or better air cooler on the AMD version anyway - who wants to wait, right?

Here's another Navi Cooler replacement, it's in German, perhaps there will be an English version, for now Youtube Google Auto-translate to English works.

He goes from just over 1800mhz @ 180 watts under load to 2040mhz @ 265w under load during brief testing. I would have liked to have seen at least a little comparison benchmarking, but still it's a nice 3 fan conversion. Maybe he will come back with a full comparison when he perfects the cooling.

AMD RX 5700 XT | Custom cooler | 2 GHz + | quiet and cool? | Overclocking
PC Games Hardware
Published on Jul 10, 2019
The DHE cooler of the new Navi graphics cards is relatively loud and does not offer very good cooling performance. We want to counteract this problem with an Arctic Accelero Extreme IV retrofit student. In addition to the radiator conversion, we show the temperatures and possible maximum cycle rates that can be achieved by the improved cooling performance.

 

yea prob would happen. for them to claim "the first to make thinnest 8 core or 16 core notebook" w/e marketing bs they throw at it. all i care is gaming laptop and workstation and a heatsink similar size to clevo P870TM would be able to cool ryzen's 8 core with almost silent fan.

it'll be nice for a 4.1 to 4.2ghz 16 cores laptop at 1.25v for workstation purposes

 

AMD: Ryzen 3000-Series CPUs Lack Manual Overclocking Headroom (tomshardware.com)

"Robert Hallock, Senior Technical Marketing Manager at AMD, has explained on Reddit (and copied below) that AMD pretty much squeezes all the performance out of every single Ryzen 3000-series processor, leaving little headroom for manual overclocking"

http://forum.notebookreview.com/thr...ers-welcome-too.810490/page-619#post-10931494

 

Oof.

Rumour: AMD Radeon VII has been declared end of life

https://hexus.net/business/news/components/132671-rumour-amd-radeon-vii-declared-end-life

 

That didn't take too long... power play registry mod for increasing OC headroom for the RX 5700XT, now performing at Radeon VII levels in 3dmark, and with a new cooler - or an AIB board - even higher performance might be seen.

SoftPowerPlayTables Unlock Navi, this changes everything!
Timmy Joe PC Tech
Published on Jul 12, 2019
Check him on youtube (German)
Original Article (German)


2.2 GHz and more | AMD Radeon RX 5700 XT Overclocked | Navi without OC limit | Exclusive Power Mod!
igorsLAB
Published on Jul 11, 2019
Free ride for everyone! We still know the SoftPowerPlayTables from the Radeon RX Vega and the Radeon VII. Now these cool tweaks are also available for Navi of my homepage, so that even the smaller RX 5700 can be easily overclocked. Our forum user hellm has done a really good job here and yesterday we determined the most sensible settings for you. The download and the exact explanation are in the review on the homepage and of course the appropriate support in the forum thread of the RTG (Radeon Tweaker Group)! It's worth it, promise! Hir goes to Review and Download: https: //www.tomshw.de/2019/07/11/unge ...


RX 5700 XT power play mod overclocking for RADEON VII performance!!!!
not an apple fan
Published on Jul 12, 2019
RX 5700 XT power play mod overclocking for RADEON VII performance!!!!


The Radeon VII is rumored to be up for reassignment to the Pro lineup - or that may just be a new sku in addition to the Radeon VII, we haven't heard anything official yet from AMD.

The Radeon VII at $650 is way over priced now for it's consumer performance - it still outperforms in compute tasks + 16GB HBM2 outperforms for professional tasks, so either a price drop as a consumer board or a reclassification as a Pro board is coming.

 

It already is practically the instinct MI-50 if I remember correctly. So ending it for consumers isn't too big of a deal. It was always a stop gap due to the retape of Navi.

So there is no problem here. They are EOLing because they are clearing inventory between this and what can be repurposed for the enterprise so that they will have room for the 5800XT potentially this fall or at CES.

Now, them not fully closing the power play tables is awesome of them. Someone needs to let buildzoid know and hope GN explores it.

 

This in a nutshell.

The compute power plus HBM2 advantage has been well known for quite some time, but for my consumer level use case, the price didn’t justify the card in any way, shape or form.

Very much looking forward to my 5700xt. It will be here Monday.

Still holding off on the 3900x. I applaud AMD for giving you near maximum performance in their cpus, but the 9900k, for my needs, is still the superior chip with the OC headroom.

 

If you already have the 9900k there's no need to rush, it'll be a better ride out past the fixes for boost and firmware updates for the new x570 and the x470 motherboards, plus if you are going to upgrade from the 9900k the 3950x is a bigger boost justifying the effort.

Plenty of time to tune the 5700xt for peak performance before the 3950x arrives.

I don't understand the kerfuffle about the chipset fan.

I've always picked the motherboards that come with or that I can mount a PCH fan - being in a hot environment at times it's helpful to get airflow over the PCH heatsink. I consider it a plus. And, after all the years I've kept these running I've not had any PCH fans die. I've had case fans die - and that's very rare as well - the weirdest was a whole batch of Noctua fans dying after 2 months within a week in of each other in one build. Swapped them all with a new batch of Noctua fans and they are all still running today 10+ years later.

If I needed to I would pull the PCH armor fascia and attach a new fan on top of the heatsink, even replace the heatsink if needed.

To those who care, my X570 chipset fan noise is a non-issue (Gigabyte Aorus Master)
Submitted 6 hours ago by Msuix Ryzen 3900X / X570 Aorus Master / Strix 1080Ti
https://www.reddit.com/r/Amd/comments/ccd4q5/to_those_who_care_my_x570_chipset_fan_noise_is_a/

"Once updated to the BIOS revision that gives PCH fan control, setting it to silent yielded the following:

0 RPM until 60c.
~1500rpm never really getting above 70c.

I run an extremely silent PC and am bothered by most noises, so I was naturally hesitant when I saw the chipset fans on the new line of motherboards. I picked up an X470 C7H just in case, and it looks like i'll be returning that.
My case fans are Noctua running 535rpm, cpu sits at mostly 600rpm-1000rpm (gaming load - noctua NH-D15). The loudest thing in my case is a magnetic storage drive that will be removed shortly. The PC is as dead silent as it was with my previous Z270-E / 7700k. No concern.

I can't speak to the longevity of the chipset fans, but if its not even spinning for the vast majority of the time I can't imagine we'll see many failures during the working lifetime of the board.

In case you were concerned, I am no longer."

This is helpful, as it looks like 3800mhz memory, maybe detuned between 3700mhz and 3800mhz with additional tweaks is the peak performance possible:

Some DDR4 3800 1:1 performance numbers on Zen 2
Submitted 2 hours ago * by Monstru501
https://www.reddit.com/r/Amd/comments/ccfw74/some_ddr4_3800_11_performance_numbers_on_zen_2/

 

Won't help much if the OEM's cripple their products...


Lenovo laptops with the Ryzen 7 3700U run 58 percent faster than HP laptops with the exact same processor notebookcheck.net

 

Although I did not see 58% performance there is some. This is also laptop performance and 14nm.

 

14nm or coming 7nm arch doesn't matter. The OEMs will continue run their own game. Some cripple their products harder than others. AMD or Intel doesn't matter.

------------------------------------------------------------

No, AMD Still Isn’t Enabling PCIe 4 On 300/400 Series Boards anandtech.com
So, to get right to the heart of matters, I reached out to AMD PR this evening to find out what’s going on with PCIe 4 support. The short version then is that no, AMD’s plans have not changed: PCIe 4 support will be disabled in the shipping AGESA for these boards.

Our plan is unchanged. For the reliability and consistency reasons cited at Computex, we still intend to disable PCIe Gen 4 for pre-X570 motherboards. That AGESA is being released to motherboard manufacturers soon.​

As things stand, any boards that currently support the feature would be using pre-release AGESAs, and as we’ve seen with our own BIOS issues , the Ryzen 3000 BIOS situation is still evolving fast. So with AMD intending to permanently disable the feature – and prevent any workarounds – AMD’s goals haven’t wavered. At best, the few boards that have beta BIOSes with the feature will lose them in the future, unless users opted to stick with an unsupported (and almost certainly buggy) BIOS.

Of course... AMD want some money from that new boards

 

Here's a good answer from the comments section that says it plainly:

Jansen - Saturday, July 13, 2019 link
" The old school perception that AMD has been fighting for 30 years is that AMD has inferior products and that it's not as reliable as Intel.

In order to fight that perception, they need to have products that are consistently faster, cheaper, and more reliable right out of the box.

Having even a few models of older motherboards not work with PCIe 4.0 would muddy the waters and cast doubt once again. It would cause confusion in the marketplace and play right into Intel's hands.

Really, AMD is trying to support its customer base by keeping the socket the same. Sure, it would be nice to have PCIe 4, but the new Radeon RX 5700 will still work fine on PCIe 3.

If you really want PCIe 4 you can always sell your old board and buy a new one, but I suspect most people will be ok just keeping their old one."

AMD makes money selling x470 chipsets and x570 chipsets, and makers are still building both, so AMD shouldn't care about whether someone wants to buy an x470 or x570,

AMD doesn't want to give a mixed message about x470. AMD stating that all x470's will not support PCIE 4.0 is clear.

Wait for cheaper PCIE 4.0 motherboard models to release. The A520, B550, and cheaper x570 motherboards.

The high priced x570 motherboards probably won't sell through as fast as the inexpensive models, so those might get marked down to sell through on Black Friday, Cyber Monday, Holiday sales too.

 

Will Asus put the middle finger up in the face to AMD or will they be put in place instead?


Asus X470 & B450 motherboards get PCIe 4.0 per update notebookcheck.com | July 14, 2019

After the cancellation of AMD you can count on PCIe 4.0 for older motherboards. In fact, Asus has released a list of models that partly support full bandwidth for both graphics cards and SSDs.

The list published in Asia (via MyDrivers ) contains a lot of cheap B450 and X470 motherboards. After the public cancellation of AMD at least Asus brings the new PCIe standard so on older boards. It uses the integrated I / O chip of the Ryzen 3000 CPUs, which provide at least 16 PCIe 4.0 lanes for graphics cards and four more for SSDs.

Or is it ok with "partly support full bandwidth for both graphics cards and SSDs"

------------------------------------------------

AMD Starts Direct CPU Sales With the Ryzen 9 3900X Tomshardware.com

AMD is cutting out the middlemen (kinda). The company used to limit direct-to-consumer sales to limited edition products, but that changed with the Radeon VII's introduction earlier this year, and now it's started to sell the Ryzen 9 3900X on its website as well. Or at least it's getting ready to--right now clicking the giant red "Add to Cart" button results in a message saying "the AMD product you are interested in purchasing is currently out of stock."

There is still the potential for direct sales there, however, especially since AMD offers the option of being notified when the product is restocked. At that point, AMD will offer direct sales for products in both of its most important product categories. The company hasn't completely shifted to a direct sales model (and we doubt it would), but it seems to have finally grown comfortable with the idea of its products bypassing retailers on their way to end users.

For AMD, the most obvious benefit of switching to direct sales would be higher margins. Retailers expect to make some money on the products they sell, of course, so they only pay manufacturers a fraction of the product's final price. That means the products have to make a profit twice: once for the manufacturer and once for the retailer. Tech enthusiasts aren't really AMD's customers; sellers like Amazon and Newegg and their counterparts are.

Offering direct sales would change that. Enthusiasts would actually be AMD's customers at that point, which might make it easier to get support whenever something goes wrong, but that's just a potential secondary benefit. We suspect that AMD's primary motivation in selling directly to consumers would be the increased profits it could make from each sale now that it wouldn't have to split anything with another company. (To put it simply: cha-ching.)

 

It is not that hard for them too make a bios switch with default off and explain the feature is beta ,etc..

 

You didn't go to AMD's site to look before posting did you?

AMD isn't selling the 3900x direct. That link that toms posted is the "how to find where to buy" page, and AMD isn't listed as a source / seller.

Here is AMD's 3900x page, and there is no AMD hosted purchasing option, it's not listed in their shop for sale by AMD either:
https://www.amd.com/en/products/cpu/amd-ryzen-9-3900x

The "Shop Now" button goes to the "Where to Buy - AMD Ryzen™ 3000 Series Processors":
https://www.amd.com/en/where-to-buy/promotions2

And, AMD isn't listed as a seller.

Vendors advertising the PCIE 4.0 compatibility of their previous generation motherboards is old info, none of them will be supported by AMD Agesa updates - and vendors need those updates for bug fixes - so unless you want to skip updates and stay on an old BIOS that still allows PCIE 4.0 transfer rates for PCIE 4.0 SSD's, you'll want to upgrade.

I wouldn't have expected PCIE 3.0 motherboards to support PCIE 4.0 speeds, and it's too bad that some vendors thought it was a good idea to try to support it - it lead to confusion even before the release.

Like I suggested earlier, wait for the A520, B550, and cheaper x570 motherboards to come out. Right now the vendors are more interested in selling high priced AMD motherboards to the same crowd that are used to spending big $ for high priced Intel motherboards.

AMD gets the same $ for their chipsets whether the chipsets are used on a budget board or a high priced board, it's the vendors / makers of the motherboards that are cashing in on Intel owners switching to AMD.