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

Ah but you are comparing a desktop ryzen that has infinity fabric and two CCX. Oh and the TDP is 65W for Ryzen 1400 for example (if that's the one you mean).
So, for the APU which will be using a Vega IGP and will have same 4 cores and 8 threads, I think the best/likely choice here for AMD would be a single CCX which could actually provide better overall performance.

Also, look at Bristol Ridge... that is an excavator based CPU with an IGP, and brings a total TDP to 65W.
Now, Ryzen is made on a smaller manuf. process, and will have Vega in it (which will probably be more efficient than the previous IGP's).

3 GhZ baseline clocks are likely what we can expect with Raven Ridge, though not 100% sure.
Also, it depends on whether Raven Ridge will be using 14nm or 14nm+... if its 14nm+, then we could expect higher frequencies and improved efficiency.

P.S. How are Ryzen chips much less efficient than Intel? Everything we saw to date indicates otherwise.

 

That leaked benchmark for the Ryzen 2500U is an ULV chip (15W equivalent) or the HQ (45W) equivalent? If it is equivalent to the intel´s U chips, then the lower base clock is in line to what is possible at such a low TDP and 4 core, as in the new quad core U chips by intel. Either way, mobile ryzen is looking promising, at least.

 

But, the real question is, will it outperform the Nvidia 1080ti?!!

 

Is it?
The U mark doesn't necessarily mean its ultra low powered for AMD.

 

They only following naming scheme similar to Intel’s Core-U serie Interpret it as you want.

 

http://markets.businessinsider.com/...rrent-generation-Platform-features-1002388669

https://www.globalfoundries.com/new...-technology-for-high-performance-applications

Looks like the Ryzen refresh will be on 12LP instead of 14nm++. This has implications for Vega as well, with the refresh being on 12nm just like Volta will be. Papermaster said "We are pleased to extend our longstanding relationship with GLOBALFOUNDRIES as a lead customer for their new 12LP technology. Our deep collaboration with GF has helped AMD bring a set of leadership high-performance products to market in 2017 using 14nm FinFET technology. We plan to introduce new client and graphics products based on GF’s 12nm process technology in 2018 as a part of our focus on accelerating our product and technology momentum." This also means the 2000 chips may be on this process.

I would also like to point out the LP nomenclature matches the nomenclature found for the 7nm process, 7LP, not the 14LPP that was used for the current 14nm Ryzen and Vega chips on the market. This may suggest that this is the IBM process on 12nm before going to 7nm next year. If so, there is better reason to believe in performance improvement.
https://www.globalfoundries.com/sites/default/files/product-briefs/7lp-product-brief.pdf
https://www.globalfoundries.com/sites/default/files/product-briefs/product-brief-14lpp.pdf

"The new 12LP technology provides as much as a 15 percent improvement in circuit density and more than a 10 percent improvement in performance over 16/14nm FinFET solutions on the market today."

@Papusan @Deks @hmscott @Rage Set @TANWare

 

Man, this computer assembly is a slow process.

in the case so far;
PSU
Main board and CPU
3tb hdd
BR drive
Card reader

The CPU is a royal pain to get those three torque screws to grab.

 

Just saw the news on tomshardware for that.
But isn't 12LP signifying low power as well?
This would indicate the process is (yet again) not suitable for high clock speeds... and will impair efficiency if people or manufacturers try to overclock or try to put out products with too high clock rates.

Vega will probably experience a drop in power consumption at current clocks... but AMD would be better off trying to integrate auto voltage tuning to optimize efficiency and also increase HBM speeds as vega seems more bandwidth starved.
Current Ryzen iteration can only benefit from higher clock speeds (for people who like single threaded programs) and software optimizations that takes advantage of the architecture (which showed that even at lower single core settings, it can be competitive to Intels higher clocks).

 

Am I the only one that didn't have an issue getting them to grab and torque down?

 

No, they changed it so that LP stands for like Leading Process, not low power. Read the 7LP link I gave.

Edit: Leading performance! http://techreport.com/news/32110/globalfoundries-fires-up-its-7-nm-leading-performance-forges

 

Hahahahaha. Welcome to the club. Yes, AMD requires you to nearly break the mobo in half to tighten the CPU bracket. It's in the manual,

 

That's the thing, we can't interpret this individually unless you want heavy bias at play.
We need more information (such as the chip's TDP).

The i7 7500U scores in geekbench:
https://ark.intel.com/products/95451/Intel-Core-i7-7500U-Processor-4M-Cache-up-to-3_50-GHz-

Now, if Ryzen 2500U is indeed a low power unit at 15W, and leaked geekbench scores real... then it would easily beat i7 7500 while being set to 2.0 Ghz... but of course, it would depend on what the TDP for this thing is.
The geekbench scores for Carrizo APU's are nowhere near this level...

Kaveri had a configurable TDP (same with Carrizo)... and A10 7800 for example scores 90% lower in multithreading... but it's TDP and clock frequencies are 65W (I think) and 3.5GhZ base clock
Carrizo to my knowledge... we have again 65W APU's for comparative purposes, but Athlon X4 845 scores close to those leaked Raven Ridge scores (about 1000 points less for multithreading it would seem).

So, what are we comparing to... which TDP... what is the TDP of this new APU supposed to be?
If its 15W, then great, AMD could smash Intel's current low power lineup easily enough... however, we don't know that.
The Carrizo A10 8700P for example scores highest about 4465 points (multithreaded) in geekbench 4.
And this could be the 35W part, but I somehow doubt it as the power supply for the laptop that achieved it was 45W - and I don't know if 15W is enough for the rest of the system if the APU alone is taking in 35W (probably over that if under load).
http://browser.geekbench.com/geekbench3/search?q=AMD+Carrizo

I could be wrong.
We need proper context and comparison data.

 

I linked a vid from I think HardOCP with tips on how to do that exact process. Must be a number of pages back.

 

Ah, so would 12nm in that case indicates an upgrade in speed and efficiency that would be similar to 7nm before actually jumping to 7nm.
If so, that would be pretty amazing and a really good thing for AMD.
It would indicate the GLOFO might not be limited to lower clock speeds anymore and AMD could clock Ryzen much higher (to or past 4.0 GhZ) with current efficiency.

This indicates that 14nm+ is no more and 12nm is the Ryzen + and Vega + (essentially).
Dangit... should I hold off on my Asus ROG 1700 and RX 580 in that case and wait for 12nm instead?
Nah... I think I can manage just fine - if I save up enough cash I can just as easily buy the next iteration when its released, or just wait for Ryzen 2 to upgrade.

I wonder what 12nm will do for Vega though.
Do you happen to know how that process compares to the one used by Nvidia (which I think is the samsung's manuf. 16nm process)?

 

Tom's Hardware Updates Best Gaming CPUs of 2017 - Recommends Ryzen in 4/6 price brackets
https://www.reddit.com/r/Amd/comments/714hv2/toms_hardware_updates_best_gaming_cpus_of_2017/

Best CPU's
http://www.tomshardware.com/reviews/best-cpus,3986.html#xtor=RSS-100

"If you don’t have the time to research benchmarks, or if you don’t feel confident enough in your ability to pick the right processor for your next gaming machine, fear not. We at Tom’s Hardware have come to your aid with a simple list of the best gaming CPUs offered for the money.

This list is for gamers who want to get the most for their money. If you don’t play games, then the CPUs on this list may not be suitable for your particular needs. The criteria to get on this list are a mixture of price and performance. Cost and availability change on a daily basis.

We can’t offer up-to-the-minute accurate pricing information in the text, but the prices in green are current. The chart is based on MSRP, while the list is based on the best US prices from Amazon, Newegg, and others. In other countries or at retail, your mileage will most certainly vary. Of course, these are new, retail CPU prices - we do not list used or OEM CPUs."

 

So, first, let's calm down. It is a 15% increase in logic density and 10% performance increase in transistor performance. This is from node/process switch compared to the 14nm. This does not include architectural tweaks which may effect IPC. You should not compare it to 7nm at this time.

Nvidia will also use 12nm on Volta next year from TSMC.

Meanwhile, Vega 11 and Vega 20 may use the new process. Depending on when release is, a vega 11 on 12nm replacing the 580 would be a better deal. But if you need it now...

On Ryzen, if you buy now, wait until zen2 to upgrade, not the refresh. More jump and fiscally sound. But, I'd recommend the 1900x or wait for the 1920 (or potentially 1900). This is truly the best deal, as the X399 platform seems more stable than the X370 platform, plus you have more pcie and quad channel memory. Use your best judgement.

 

I'm not trying to compare it to 7nm... I'm just asking is the 12nm a precursor to IBM's 7nm?
Namely, how much of a connection does the 12nm process have to IBM's 7nm?

I know that Nvidia will use 12nm on Volta from TSMC... any comparatives available between that manuf. process and the 12nm LP AMD will use?

Yeah I was thinking that if I buy now, I'll wait until Ryzen 2 or Ryzen 3 for an upgrade, because as you say, it would result to a much nicer jump.
I'm still gonna get a ridiculous jump over my current laptop in specs, performance and efficiency.
But note that I'm getting the Asus ROG all AMD laptop (Ryzen 1700 cpu and rx 580)... not the desktop platform.
Anything past 8 cores even on the desktop is way beyond my budget.

 

Still debating that on other forums. No one knows yet. Same with the tsmc comparison. The process was announced this morning, so everyone is in the dark speculating.

 

More news, evidently GloFo announced an IBM 14nm SOI finFET today as well. Now, the 7nm is not a SOI finFET, just a finFET, so it doesn't confirm or reject this being the 7nm on a larger node.
https://www.globalfoundries.com/new...custom-14nm-finfet-technology-for-ibm-systems

 

AMD Announces 2nd Gen Ryzen & Vega Coming On 12nm Next Year
http://wccftech.com/amd-announces-2nd-gen-ryzen-vega-launching-12nm-2018/

 

Noctua Announces Two AM4-compatible Low-profile Coolers
https://www.techpowerup.com/237174/noctua-announces-two-am4-compatible-low-profile-coolers

Spoiler: Details...

"Noctua today presented two updates to its range of quiet low-profile CPU coolers. Successors to the award-winning NH-L9a and NH-L12, the new NH-L9a-AM4 and NH-L12S models now support the AM4 socket of AMD's Ryzen processors. Whereas the NH-L9a-AM4 stands a mere 37mm tall, which makes it ideal for use in compact HTPCs or Small Form Factor cases that offer minimum clearance for CPU coolers, the slightly bigger NH-L12S comes equipped with the new NF-A12x15 PWM fan and combines excellent compatibility, surprising efficiency and remarkable flexibility.

"AMD has made a big impact on the market with its Ryzen architecture and the platform will become even more interesting for compact HTPC builds when the first Ryzen-based APUs are introduced next year," says Roland Mossig (Noctua CEO). "We're thus updating two of our award-winning low-profile models to support Ryzen's AM4 socket.""

 

Custom-design Radeon RX Vega Cards by Mid-October
https://www.techpowerup.com/237156/custom-design-radeon-rx-vega-cards-by-mid-october

"Still reeling under supply issues and overpricing, AMD's Radeon RX Vega line of graphics cards may finally be available in custom-design products from the company's AIB (add-in board) partners by mid-October, according to a Hardware.fr report. ASUS was the first to announce custom-design RX Vega 64 and RX Vega 56 graphics cards under its ROG Strix series, back in August. The cards were, however, nowhere to be found in the markets.

AIB partners will begin announcing their custom-design RX Vega series products in the coming weeks, with retail availability slated for mid-October. Radeon RX Vega 64 is currently available in three AMD-reference design SKUs, the standard reference-design, the premium "silver" air-cooled reference-design, which features a brushed aluminium cooler shroud and LED ornaments; and a more premium AIO liquid-cooled variant with higher clocks. The RX Vega 56 is available in vanilla standard reference-design."

For when the RX Vega Custom? October?
http://www.hardware.fr/news/15240/quand-rx-vega-custom-octobre.html

 

AMD To Change Suppliers for Vega 20 GPUs on 7nm, HBM2 Packaging for Vega 11
https://www.techpowerup.com/236874/...ega-20-gpus-on-7nm-hbm2-packaging-for-vega-11

Spoiler: Details

"AMD's RX Vega supply has seen exceedingly limited quantities available since launch. This has been due to a number of reasons, though the two foremost that have been reported are: increased demand from cryptocurrency miners, who are looking towards maximizing their single node hashrate density through Vega's promising mining capabilities; and yield issues with AMD's Vega 10 HBM2 packaging partner, Advanced Semiconductor Engineering (ASE). It's expected that chip yield for Vega 10 is also lower per se, due to it having a 484 mm² die, which is more prone to defects than a smaller one, thus reducing the amount of fully-enabled GPUs.

AMD's production partner, GlobalFoundries, has historically been at the center of considerations on AMD's yield problems. That GlobalFoundries is seemingly doing a good job with Ryzen may not be much to say: those chips have incredibly small die sizes (192 mm²) for their number of cores. It seems that Global Foundries only hits problems with increased die sizes and complexity (which is, unfortunately for AMD, where it matters most)."

"Due to these factors, it seems that AMD is looking to change manufacturers for both their chip yield issues, and packaging yield problems. ASE, which has seen a 10% revenue increase for the month of August (not coincidentally, the month that has seen AMD's RX Vega release) is reportedly being put in charge of a much smaller number of packaging orders, with Siliconware Precision Industries (SPIL), who has already taken on some Vega 10 packaging orders of its own, being the one to receive the bulk of Vega 11 orders.

Vega 11 is expected to be the mainstream version of the Vega architecture, replacing Polaris' RX 500 series. Reports peg Vega 11 as also including HBM2 memory in their design instead of GDDR5 memory. Considering AMD's HBM memory history with both the original Fury and and now RX Vega, as well as the much increased cost of HBM2's implementation versus a more conventional GDDR memory subsystem, this editor reserves itself the right to be extremely skeptical that this is true. If it's indeed true, and Vega 11 indeed does introduce HBM2 memory to the mainstream GPU market, then... We'll talk when (if) we get there.

As to its die yield issues, AMD is reported to be changing their main supplier for their 7 nm AI-geared Vega 20 from GlobalFoundries to Taiwan Semiconductor Manufacturing Company (TSMC), who has already secured orders for AI chips from NVIDIA and Google. TSMC's 7nm and CoWoS (chip-on-wafer-on-substrate) capabilities have apparently proven themselves enough for AMD to change manufacturers. How this will affect AMD and GlobalFoundries' Wafer Agreement remains to be seen, but we expect AMD will be letting go of some additional payments GlobalFoundries' way."
Sources: DigiTimes on Vega 11, DigiTimes ASE, DigiTimes on vega Shortages

 

AMD's James Prior Clarifies Threadripper's "Dummy Dies"
https://www.techpowerup.com/237139/amds-james-prior-clarifies-threadrippers-dummy-dies

Spoiler: Detail...

"Much has been said regarding AMD's Threadripper CPUs, particularly when it comes to how they are manufactured. At first, we thought Threadripper was actually EPYC in disguise, due it having what appeared to be four full-fledged 8-core modules - the same design as AMD's server-bound 32-core EPYC chips. The presence of gold-plating under all four dies seemed to confirm that these were in fact four full Threadripper dies, instead of two dies and two spacers (as AMD's statements led us to believe) for even IHS pressure on the four dies, instead of the uneven pressure that would result from the chip only having two physical dies present.

It seems the truth is, as always, somewhere in-between. On Twitter, James Prior came forward to shed some light on the issue, clarifying what exactly is going on and justifying AMD's usage of "dummy" and "inactive" nomenclatures. First up, no, readers: there is virtually no way to reactivate those unused dies. As James Prior himself said, "Threadripper is not a Epyc processor. Different substrate, different dies. 2 dies work, other 2 have no path to operation. Basically rocks." Those are some expensive pieces of rock, for sure. However, AMD's choosing of "dummy" instead of inactive seems correct here: "(...) exactly why they're not described as inactive, but dummy. Doesn't matter if they were dead, or active, they're not going to work.""

 

$6,000 AMD THREADRIPPER BUILD... INSANE!
Note: Asus AI Suite version caused massive dropped frames - Rant Starts @ 17:20

ASUS ROG ZENITH EXTREME X399 Build - Ryzen Threadripper 1950X / GTX 1080 Ti (2-Way SLI)

THE ULTIMATE AMD PC BUILD 2017 - Threadripper AND Vega!

ROG Zenith Extreme Gaming Motherboard Overview

 

AND Vega RX64 LC & RX56 vs. Aorus WaterForce 1080ti

 

RYZEN 3 1200 + GTX1060 3GB: 1080p Gaming Benchmarks!

 

https://www.tweaktown.com/news/5921...signare-ex-motherboard-looks-great/index.html

 

Mine lives. 4.0 GHz at 1.350 first try. The memory was a pain as had to get it OC'd as dual channel and the one stick ws a pain to get it to recognize and seat properly. In the end 3466 memory too.

 

I'm still waiting for my custom loop to have that fun...

 

The Enermax is awesome, Prime 95 for 30 minutes HWmonitor showed peak of 295W in the package and a top of 77C. at 4.0 GHz though idle is almost 50w to the CPU. R15=3487 so I am pretty happy at the moment.

I should note at first OC of the CPU I had 4.0 GHz but only 2166 memory, it refused to go further. At that R15=3075 best run.

 

Oh wait... that's the article in question about AMD changing suppliers for 7nm...
Hm...
But yes, reading it again, it seems like the 7nm node change indicates a switch for GPU... not Ryzen.
This is probably because they found issues with GLOFO meeting demand and not being able to produce larger sized chips, and of course HBM delays from another company.
At any rate, not technically AMD's fault.
Vega seems like an efficient architecture when undervolted, but we know that most people don't care about the details and reviewers won't pay almost any attention to that.
For crying out loud, I was called an 'AMD fanboy' for presenting facts on undervolting Vega 56 and how it can beat 1080 when undervolted and overclocked on both core and VRAM for a lower power consumption than 1080.

So, sticking with GloFo for CPU might be viable for them, as IBM's manuf. process is probably going to be superior on 7nm.
Now, for the GPU, TSMC might be a better option as that will (finally) give us a better way to compare the competing GPU's on similar clocks and power efficiency out the door.
I just hope AMD can get voltage tuning sorted out the door similar to Nvidia.

I'm more than happy to undervolt and get AMD as I know what it can do, but it doesn't do much for their overall 'image' with some other people.

 

It sounds like you have a really good chip. I hope you try to push to 4.1 or at least 4.075 @1.385.

 

My concern is if I have to reset cmos and get 3466 back. Took a while to get it just right.

 

I always take copious notes to catch the details as I set them and keep running lists of the ones that seem to work and work best.

I hate it when I run till I am tired and then neglect to write the current settings down before starting up again, only to loose a couple of hours work.

Even with MB's that keep multiple profiles you can get caught without an in firmware backup.

 

Actually I do not have the gold I thought I did. I ran Prime95 again and after 5 minutes a couple of workers error out. I had to got to 1.3625 for 100% stability. Still a happy little clam here.

 

SL binned [email protected] is 1.325v

 

It's nice to know SL is conservative in their voltages... I wonder what they use for 4.0ghz, or above...

 

I can run at 1.325 but again prime 95 has stopped workers. The other issue is cooling. With just the stock old water AIO I do not think I would be able to get 4.0 GHz. This could be the reason SL is conservative with voltages too.

 

The SL binning includes P95 stability at [email protected]. That is the one I bought.

But as TANWare mentioned, the heat. This is why I'm at 3.85 until that loop is together. Then I'll give some information on how well the binned chips do and see how good my chip is.

 

I tried 4.1 at 1.3725, boots but Cinebench locked up. At 4.050 @ 1.3725 R15 = 3523, did not try Prime95 though.

An Edit; now that I have found a socket that the onebad memory stick seems happy with I can set the XMP without an issue.

 

Regarding SL's testing methodology...

 

The changes in testing and warranty is what added to the reasons to purchase a top bin. We've discussed the percentages previously (no longer public), but the additional testing means more likely to be good at said voltages. You still get the ranges, and the top bin only shows the ceiling on voltage at a set multiplier, not the floor.

Also, the large box of items just got here. Now, need to have the cpu block ship...

 

well, https://www.3dmark.com/3dmv/5644982
https://www.3dmark.com/3dm06/17996515
https://www.3dmark.com/3dm/22292375

 

https://www.tweaktown.com/news/59231/amd-radeon-rx-vega-32-28-rumored-coming/index.html
News on the Polaris replacements.

 

AMD Radeon phases out the CrossFire brand as multi-GPU gets more complicated-Pcworld.com

"When AMD released the Radeon Software 17.9.2 drivers that enabled support for multi-GPU solutions in Radeon RX Vega cards this morning, the announcement post didn’t mention CrossFire—AMD’s longstanding brand for multi-GPU (mGPU) solutions—a single time. An AMD representative revealed why in response to a PCWorld query: In this complicated mGPU age, Radeon is abandoning the CrossFire brand."

 

Unfortunately, this falls as the 1070 Ti MSRP was announced in a rumor (not a coincidence): $429 or around that.

Forcing sli through profiles is favorite for some. Also, this is cut to just below the size of a 1080, but only uses gddr5, not a variant. So Vega 56 did have a reaction.

 

https://www.3dmark.com/hall-of-fame-2/fire+strike+physics+score/version+1.1

Number 27 in the 3DMark Physics Hall of Fame!!! 30,736

I saw my username in this video towards the end, so I made it a mission to pass this Youtuber (I was originally number 34).

 

Don't worry, in two to three weeks, you'll be one spot lower most likely! ;-)

 

Using Overclocked 4.25ghz AMD Threadripper to TOP BENCHMARK LISTS!