Ryzen vs i7 (Mainstream); Threadripper vs i9 (HEDT); X299 vs X399/TRX40; Xeon vs Epyc

AMD needs to get rid of what they already have in stock/shelves. Not fun sell today’s chips with loss. See... Nvidia and 1060/2060 situation.

 

I personally think there will be announcements, releases, and details.

To be clear, the releases will be like the APUs built on 12nm, which will come with details, and maybe the low power 3000u products.

7nm products will be announced and a little more detail given, like IPC or a demo. I'm thinking like what was shown at new horizons Dec. 2016.

On GPUs, AMD has dropped prices and created a plan to get rid of them over 3-4Q, announced in Q3 earnings. But unlike Nvidia, they gave clear hedging language, did not have the puffery or potentially actionable language surrounding the GPUs, etc.

On CPUs, that makes no sense whatsoever. In GPU and desktop CPU segment, the reason margins went from 36% to 39% was because CPUs went from around 50% of sales in that segment to 70% on sales, which hits to the heart of exactly how much margin they already have built in.

Moreover, did you see the CPU sales in November and December? I know you did, as you tried to use those sales as proof AMD was scared of Intel and their CPUs weren't worth it, blah blah blah.

Instead, they were likely clearing the inventory channels. You saw gen 1 Zen 1700 for $150 at one point and 1700x between $180 and 220 at its lowest, the 2700X was as low as $250 (down from $310 at the time, meaning 20% discount), although $270 was more commonly seen, etc.

Also, AMD went through this with the first crypto boom, Nvidia did not (because Nvidia's compute was not as good back then). Nvidia made no accommodations for the used market, overproduced, then gave ****ty prices on new products to try to move the old, but really just pissed off consumers.

Believe it or not, they know how to do this. They've done it for years and had Carrizo for god sakes. You think they don't know how to deal with aging inventory after the first crypto boom and Carrizo? Please...

Sent from my SM-G900P using Tapatalk

 

Well I guess we will know soon enough and either HWU or other individuals will be wrong. Hoping AMD announces something good though. Intel needs that kick in the rear. The last couple years have been excellent in the CPU sector because of competition (again) and I hope it doesn't slow down anytime soon.

 

With the big hype around 7nm there is not fun risk sitting on aging inventory. Second best.. Throw they out with big discount but will hamper the sales of new chips. I expect you as me don't have the numbers for AMD's inventory. Could as well be late release due production problems or even build up big enough inventory of new chips. Pick what you prefer for yourself

 

Why the fick do you do this? You chose worst case to put bad thoughts on AMD in people's heads, them do the opposite with Intel and Nvidia when, right now, it's very much the reverse, except on GPUs. Corrizo was killed off this summer. They promised to have support for these chips for 3 years CPU side. That means having enough of the pro series for commercial drop in replacements.

If you want to worry about inventory, wait for the 10-K in about two to three weeks. Truth is, Intel's set to have a worse 4th quarter, then bad first quarter of 2019.

As to second fiddle, due to size before Ryzen and how much market share AMD took equaling billions, so I think they are sitting pretty, especially with excess 7nm capacity at TSMC. Can't say the same with the 14nm expansion for Intel, the shortage continuing there, etc.

As to GPUs, at the midrange, 580=1060. Simply put, they are equivalent and which wins depends on the game. I personally expect Navi in the second half, despite rumors to the contrary. That is the original delivery cadence, historically, and allows to clear inventory easily, with Q3 being 3 quarters after they said a year to absorb inventory, meaning they have a plan. And on GPUs, their Market share was in the 30s, much lower on CPU side before Ryzen.

So there isn't an actual worry as to either if you look at their financials, but you talk out of your ass general business principles without even examining what the **** a company is doing before you open your mouth. If being honest, between that and the "bro" mentality, those are my problems with you.

Edit: This also as AMD has said the 7nm process has been excellent and is exceeding expectations, meaning most likely both performance and yields. But, you don't point to real, concrete statements or evidence to do your fearmongering, to be clear. I reference information known to the market on AMD, Intel, and Nvidia to do my analyses.

https://realmoney.thestreet.com/art...gs-might-provide-hard-lesson-or-two-investors

The Big Picture
If one was bullish on AMD going into its Q3 report on a belief that its 7nm products can drive large share gains in 2019 and beyond, the company's Q4 guidance and commentary shouldn't do much to make one question that thesis. Likewise, if one was bearish on Intel ahead of its Q3 report due to worries about its 10nm issues and/or out of a belief that DCG's end-markets will slow, its results and guidance don't really put a dent into either thesis.

Nonetheless, AMD plunged on news that near-term demand for some of its existing products will be weaker than expected. And Intel rallied on news that businesses that had already seen demand improve in 2018 had seen it improve some more.

Perhaps the key lesson here for tech investors is to stay mindful of near-term business swings -- and how markets could react to them -- even if one is convinced that new products and technologies will have a big impact on a company's fortunes farther down the line. That might doubly hold in a risk-sensitive market environment like this one.

Sent from my SM-G900P using Tapatalk

 

@jclausius @Papusan @Mr. Fox @hmscott @jaybee83
So, after seeing what Intel is doing with their 3D stacking, I feel I must eat my words
Intel has moved the cache to the active interposer with some or all of the I/O elements. Putting the cache on the interposer takes more engineering to keep the latencies down
For that, I am extremely impressed

I still want more details, but it is likely a shared L3 that feeds upwards to both the main CPU cores and the small Atom cores. Also, with them allotting more space to the GPU, I wonder if Intel is exploring a shared caching system between the CPU and GPU cores, which would actually be amazing because it could quickly split the physics workload to the CPU with the other workloads to the GPU, which could have aided in the numbers we are seeing on the iGP front. Shared caching should be saw as almost a unified memory architecture, except bringing it to the fast bandwidths found within a caching system. There is so much more I want to know!


Edit: On further thinking, you'd need a much larger L3 victim cache design if including the iGP in at this round, but it may come in the future. But, the shared eDRAM is very likely (which is like an L4 cache) and the L3 is likely shared between the little.BIG design of the CPUs, which is still very impressive.
https://www.pcgamesn.com/intel/intel-lakefield-project-athena-mobile-pc

 

Intel Core i9-9990XE Basin Falls CPU Rumored With 14 Cores And 5GHz Boost Clock Hothardware.com | Jan 15, 2019
Intel is rumored to be releasing another Core i9 Extreme processor, though the new SKU, if it does in fact exist, is a bit of an oddity. For one thing, the upcoming Core i9-9990XE will not be sold through retail channels, and will instead only be made available to system integrators through a closed auction...

 

Are people gonna want this CPU though, I imagine it's gonna be expensive for the number of cores it offers. I'm thinking most people in the market for a CPU with this many cores are more about increased core count rather than high frequency. I mean gamers aren't gonna buy this, as 9900K is the optimum really, no need for more cores, plus the architecture of this CPU is not as good as the 9900K architecture when it comes to gaming. When they say it's an "oddity", I don't think they're wrong! Doesn't make sense to me, makes me think the rumour is not true.

 

too expensive, too hot. some say the mesh might be binned, i doubt it tho.

 

makes sense that they would opt for the 14 core sku though, if the rumour is true. the 14 core chip was always the one to get imho, offering best of both core count and high frequency in the HEDT lineup. oddly enough, it was never popular with the consumers, probably to do with being in that "gap" between the rather popular 8-10 core and 16-18 core models...

Sent from my Xiaomi Mi Max 2 (Oxygen) using Tapatalk

 

Sounds like Intel is trying to counter an unreleased 16-core that may hit 4.7+GHz. Especially if, since that demo read system power draw, the 8-core shown off by AMD may have been a 65W part with 6.5% IPC over Intel running 4.4GHz. (The EPYC server chips were 10-15% IPC over Epyc first gen, which was Zen 1, about 7% slower IPC than Intel, which means about 6% more than Intel next gen).

But, that aside, whether true or not, the threat of such a thing is showing Intel pull out all the stops, while at the same time showing how little they have left in the tank until 10nm chips come. And with only laptop chips coming this year and the rest of 10nm next year, between this and the rumored 10-core mainstream chip, it has shown how far things could have been pushed in a short time but for companies wanting to bleed us dry or bad management (talking both AMD and Intel, although Nvidia deserves an honorable mention in this regard, even though we are talking CPUs and they do exclusively GPUs).

Sent from my SM-G900P using Tapatalk

 

Limited amount chips. There is buyers out there who want something special like this in their special desktop builds. A few hundred a quarter, isn’t exactly the same as the anniversary edition (50.000). People buy more strange than this. But we don’t know if this is real or not. A hybrid chips which fits between gaming and work. Always customers for this.

 

@ajc9988 Is it patcher still relevant for AMD Ryzen these days? https://forums.guru3d.com/threads/i...intel-c-compiler-will-criple-your-cpu.403826/

 

Special Report: AMD’s Comeback Story - Intel vs. AMD Marketshare
Gamers Nexus
Published on Apr 11, 2019
This GN special report looks at AMD marketshare change since end of 2016 through first quarter of 2019. We use sales data from our viewers to determine sentiment shift. All charts are rounded to keep them clean, so might be +/-0.5%. Our special report looks at years of GN affiliate sales data to determine popular CPUs in the enthusiast DIY segment. Note that, as stated in the video, this is strictly based upon our own audience. This means that we can't plot the total marketshare of either company (where Intel does still hold more), but we can minimally plot buying trends among our viewers and readers. The data is also going to be skewed by our product reviews, where positive reviews are more likely to drive sales to that specific product. Beyond this, the data is interesting and shows what's happening within our market niche between AMD and Intel.


GN Special Report: Intel vs. AMD Sales Volume & AMD's Comeback Story
By Steve Burke Published April 11, 2019 at 9:44 pm
https://www.gamersnexus.net/guides/3458-gn-special-report-intel-vs-amd-marketshare-comeback-story

 

Just saw this question. Sorry for the delay. I know there were two issues back then: 1) Intel crippling on purpose, and 2) CMT optimization. For the first, that was seen more around 2008 and 2009, with the antitrust case brought in Europe ruling that Intel illegally used their compiler to gimp the competition. I think 2015 was conclusion of either the appellate case or the final, highest court decision. As such, I doubt it is much of an issue, although Intel's compiler isn't the best C++ compiler for Ryzen. AMD has made great strides in C and C++ optimizations for open source Linux compilers in recent years. But I don't think Intel gimps it like they used to because AMD and regulators are watching.

The second issue is that practically no one optimized for CMT, instead focusing on SMT/HT. This did harm performance on AMD chips, but it was a decision by software vendors. AMD doesn't use CMT now, so not an issue.

Sent from my SM-G900P using Tapatalk

 

PowerEdge Plus AMD EPYC Equals Better Server Performance in Half the Density, at Half the Cost
Zaira Noce • April 9th, 2019
https://blog.dellemc.com/en-us/powe...SM&cid=3497&lid=spr2242523616&linkId=65879945

" Dell EMC PowerEdge R6415 achieves #1 results on TPCx-HS Benchmarks for 1-TB and 10-TB scale factor

It used to be said that data runs the world. They were wrong! Data drives insights and insights are what matter. However, getting to these insights is hard, time-consuming and expensive. So much so that companies are turning to trusted hardware vendors to find the right tools and solutions to accomplish this very fundamental task. But figuring out which tools you need to uncover those hidden gems can be a challenge. What do most customers do to evaluate the right platform for their projects? They look to industry-standard benchmarks for guidance.

One such benchmark, TPCx-HS, provides guidance related to big data analytics by stressing both hardware and software stacks, including the execution engine (MapReduce or Apache Spark™) and Hadoop® Filesystem API compatible layers. It provides a vendor-neutral evaluation of both a product’s performance and price-performance ratio, thus helping customers make the right choice for their large cluster, data analysis projects.

The Dell EMC PowerEdge R6415 is a single-socket dense server that offers the right feature set for data analysis, as evidenced by its test results in the TPCx-HS benchmark.

As highlighted above, when compared to the Cisco UCS C240 M5 2U/2 socket servers, the PowerEdge R6415 single socket platform has better price/performance:

• 45% better @ 1-TB Scale Factor
• 53% better @ 10-TB Scale Factor

Here’s a comparison of the configurations used in this benchmark for the 1TB Scale Factor:

1-TB Scale Factor

The results speak for themselves. The single-socket PowerEdge R6415 is a great solution for Apache Hadoop®-based, big data environments without unnecessary and extra costs. Similarly, this cost-effective server is also ideal for virtualization, scale-out and software-defined storage, bringing high-density computing to the data center.

The PowerEdge R6415 delivers the following high-level key server functionalities:

• Lower TCO delivering capabilities often requiring dual socket servers
• Over 3x more direct PCIe 3.0 NVMe* for high bandwidth, low latency storage access
• High core counts optimized for high performance, multithread architecture workloads

Dell EMC servers help you prepare for innovation in your business and can help you solve your IT challenges. Learn how to turn data into insights. Follow @DellEMCServers and visit dellemc.com/servers to get started.

*Based on Dell Internal Analyses versus comparable Dell EMC PowerEdge R630 configuration."

PowerEdge Plus AMD EPYC Equals Better Server Performance in Half the Density, at Half the Cost | Direct2DellEMC (blog.dellemc.com)
submitted 4 days ago by Aleblanco1987
https://www.reddit.com/r/hardware/comments/bbmdum/poweredge_plus_amd_epyc_equals_better_server/

 

Its still relevant, w/o Intel complier optimisation my intel cpu performs like BS on CB r11 and r15. I want Intel Optimisation trick.

 

Not from the April Fool's day.

MSI Betrays AMD's Socket AM4 Longevity Promise: No Zen2 for 300-series? Techpowerup.com | April 15, 2019

Greedy motherboard vendors such as MSI want you to buy a new motherboard every two generations of processor for no sound reason at all. MSI is reportedly blocking support for 3rd generation Ryzen "Matisse" processors on its AMD 300-series chipset motherboards, including those based on high-end AMD X370 and OC-capable B350 chipsets.

This would also put those who own $300 motherboards such as the X370 XPower out of luck. ​

To recap, AMD announced on numerous occasions that it doesn't want to be a greedy clique like its competitor, by forcing motherboard upgrades and promised that socket AM4 motherboards will be backwards and forwards compatible with at least four generations of Ryzen processors, running all the way up to 2020.

Other motherboard vendors could follow MSI's suit as a representative of another motherboard vendor, on condition of anonymity, told TechPowerUp that "Zen 2" processors have steeper electrical requirements that 300-series motherboards don't meet.

 

See, for support of pcie 4.0 and other features not being included, I'd be OK with this. Not doing it at all, if I bought MSI motherboards, I'd start rethinking that stance moving forward.

Sent from my SM-G900P using Tapatalk

 

Just hope this trash doesn't happen "Other motherboard vendors could follow MSI's suit as a representative of another motherboard vendor". Then it doesn't matter what OEM you will support. Only go with the brand whitch offer what you want.

 

Hopefully it's just bad customer support information - for MSI's sake, here's the source post of the articles online:

MSI won't support Ryzen 3000 on *any* B350/X370 boards i.redd.it
Submitted 10 hours ago by master 3553 R7 1700X | RX Vega 64
https://www.reddit.com/r/Amd/comments/bdfsd7/msi_wont_support_ryzen_3000_on_any_b350x370_boards/

thomasjjc R3 2200G / R5 2400G 192 points 9 hours ago
"I once contacted MSI about the possibility to overclock the 2400G IGPU on the MSI B350I Pro AC. The option wasn't available on that board at the time. They told me, that that functionality was unavailable and not planned for that mainboard either.

Yet they still provided the option with a later bios version.

So this response for now just means that they have a very bad customer support."

notsheldogg 91 points 8 hours ago
"In most major tech companies, the customer support/tech support teams don't get information until the official release date."

master3553 R7 1700X | RX Vega 64[ S] 51 points 8 hours ago
""We can't disclose that information at this point in time" would be an appropriate response then...I'm not saying that this isn't the case, though providing wrong information is worse than providing no information"

bakerie 29 points 6 hours ago
"This thread is so eerily similar to this one that turned out to be false:
https://www.reddit.com/r/Amd/comments/b4812j/asus_prime_b350plus_wont_receive_ryzen_3000/

master3553 R7 1700X | RX Vega 64[ S] 12 points 6 hours ago
"I'd hope that this turns out similarly"

The MSI 400-series has already received the BIOS update for Zen 2 in March, and no MSI 300-series update has shown up, but it's not unusual for older boards to lag in updates. Hopefully this is all just a misunderstanding.

 

Intel Core i5 9400F vs. AMD Ryzen 5 2600X, Six-Core Sub-$200 Gaming Battle!
Hardware Unboxed
Published on Apr 15, 2019

 

If I am forced to purchase another board I will look at it very carefully in the future and a guaranty of not getting the manufacturer to start this trend is a must!

 

AMD 50th Anniversary Ryzen 7 2700X Leaked (Updated)Tomshardwae.com | April 15, 2019 at 5:20 PM


AMD's 50th Anniversary is fast approaching on May 1, and several new postings at online retailers have revealed that AMD is on the cusp of releasing a new 50th Anniversary Ryzen 7 2700X to celebrate. (Hat tip to @Dayman58 for the spot).

The chip is listed for preorder at $340.95, which is well above the typical ~$295 street price, but that's to be expected with a collector's item. You remember @hmscott ? No need to charge more for a Anniversary item vs. the standard. And even more wasted money if it's comes without increased clock speed.

http://forum.notebookreview.com/thr...-–-6c-12t-5-1-ghz.815636/page-8#post-10748183

 

MSRP for the 2700x is $329, so a special edition 2700x that includes a fancier MAX cooler for $340 seems to be a quite normal increase to account for the cost of the cooler.
https://www.newegg.com/Product/Product.aspx?Item=N82E16819113499&Description=2700x cpu&cm_re=2700x_cpu-_-19-113-499-_-Product

As mentioned in the article there were special edition MAX cooler edition CPU's elsewhere already, so again not unusual:

" AMD has also released special edition chips in the past–the company released its Limited Edition Ryzen 7 2700 and Ryzen 5 2600X MAX bundles last year. Even though those chips, which come with a better Wraith Max cooler than the standard models, still haven't made their way to American shores, you can still find them for sale overseas."

None of that is anything like the $100 jacked up 8086k price, which traded on the promise of slightly higher OC, which was also available with new turns of the 8700k, both of which have been left behind in capability, soon to be further behind when AMD releases their new 7nm CPU's.

Which brings up the most unusual thing about this rumor of a special edition 2700x, WTH would AMD think we'd buy it? Given the release of new 7nm 3xxx CPU's announced the next day - or soon thereafter, why would anyone want a special edition 12nm CPU??

IDK, this rumor seems bogus to me.

 

The 2700X comes with a bundled Wraith Prism cooler, and it's often considerably less than $329.

Semi-OT, I was planning to wait for the summer/fall to buy a 3000-series, but my old system had...other ideas. If any of the leaked specs on the 3000-series are accurate, that will be a real beast -- a lot better deal than any "anniversary" package.

 

Jack up the price for celebrating its 50th Anniversary is exactly same as Intel did. Nothing more or less. Intel with higher clocks and AMD with a useless cooler. Can't see there is an difference in this. But I would prefer higher clocks vs the fancy cooler You mean the cooler is worth near $50 usd on top? I don't think so

 

Yeah, I agree. But, due to adoption of PCIe 4.0 and soon DDR5, plus the VRM, there may be reason to upgrade the board next year, although I'd hope to see compatibility nonetheless.

With rumors of TR3 coming in Q4 this year, it is a crap shoot on me upgrading to it or waiting for next year and going to Epyc. Granted, no one really is utilizing pcie4.0 bandwidth yet, both Intel and Nvidia are looking at server products with speeds like pcie5. But we should also get the first slot supporting PCIe 4.0.

On mainstream, I read another article that explained the problem better. Certain MFRS. Went with a 16MiB bios chip on first gen Ryzen. The second gen boards used 32MiB chips. Because of that, unless you have like a bios flashback, two bios, or similar, you would need to fork bios development for different generations.

So MB MFRS cheaping out on an aspect of the MB because they did not think AMD had anything read the market wrong and made a bad decision. Kind of sucks.

I agree with @hmscott, this product, if true, makes NO SENSE! Doesn't mean it isn't true, just who would want to buy his with the 3000 series allegedly fully launching 2 months and one week later (7/7) and with computex keynote on may 27th? This would seem like a forced release for some reason, much like the radeon vii, more than an anniversary product. Then again, they may have been honing their binning process with these chips while preparing for Ryzen 3000 and want to recoup development costs. I don't know. Or maybe they might be doing it since MB MFRS not being ready with new x570 boards in time, which we already know that is part of what AMD is waiting on to release new chips.

But either way, it just doesn't seem like the product to make a bold statement. The 8-core 95W 3XXX in a highly binned anniversary edition would make WAY more sense. Even I might have grabbed one of those (even without currently having funds to build out a whole system). Even without new boards out, many X470 boards already have the firmware to support the new chips.

But who knows. We'll find out for sure in about two wks.

Sent from my SM-G900P using Tapatalk

 

To build on the analogy, Intel released the anniversary edition about 3-4 months before their next gen, went with highly binned current gen processors, even though it only added 100-200 MHz or so. AMD is releasing current gen about 2 months before.

Either way, I don't feel either were worth buying. People will buy it, but that doesn't mean it is worth it.

Sent from my SM-G900P using Tapatalk

 

Put it the other way... If people was in the market for buying current gen chips now. It would be more wise buy highest clocked cpu vs. buy a lower clocked chips included cooler (this if they go for unlocked chips). But I'm agree with you, buy this one now with no changes in bin or clocks is stupid. Even worse if it's the same chips for higher cost but bundled with a prettier cooler.

 

I look at it this way: the 8700K was selling at $330 before that summer, and $370 during the summer the 8086K came out at $420-430. It makes no sense to pay $50-100 premium, or about 13.5-30% more for 1.9-3.8% more frequency.

In the same way, the 2700X currently sells around $290-310, with occasional deeper sales. Paying $340 is about 1/6th of that price, or about 16.7% premium, for a nice looking cooler, potentially the packaging, but no info on it being binned, etc. Once again, it makes no sense.

In both cases, I'd say buy the original, not the anniversary edition, because the price premium doesn't give the performance value with it.

I said the same thing on the 9900K. Not worth it because it is around 22% faster in productivity over the 2700X, around 11-13% better at games, but costs 63-75% or more than the 2700X, meaning you are paying a 40%+ premium for productivity or up to 50%+ premium for gaming. At $420, the premium would be only around 20% productivity and 30% for gaming. Still high, but much more reasonable.

Edit: with that said, I paid a 25% premium for a CPU that does [email protected] daily driving, which is just 4-5% over the same CPU unbinned but way better voltage curve. So a do as I say, not as I do situation.

Sent from my SM-G900P using Tapatalk

 

I'd also note that most AM4 motherboards will support Ryzen 3000 and maybe even 4000 CPUs, allowing for future upgrades without needing a new motherboard. It's a lot harder to see much of a forward roadmap for LGA1166 or whatever it is.

 

The comparison is between MSRP's, the street price's vary too much. $11 difference between $340 and $329 seems reasonable for a collectors version with an upgraded cooler.

We don't have the MSRP for the collectors version yet, it may very well be the same $329 MSRP and the $340 is a jacked up price due to the limited edition. We may need to wait for the actual announcement to know for sure.

I put in the link to newegg to show that the MSRP is $329, the street price has been arrived at over a year, the "special" edition model street price of $340 with MSRP unknown - it might even be the same $329 MSRP, we don't know yet.

Limited quantity items prices are often inflated due to perceived rarity or demand, which may turn out to not be the case with a 12nm CPU being released shortly before 7nm CPU's are released.

The difference between $329 and $340 is $11, not much of an MSRP price hike, and even less of a limited quantity price hike, so for now it's a negligible change.

Right now the 8086k is going for $589, 8700k $369, real value there. And, the 9900k is $500 - all overpriced for what they are.

No wonder AMD consumer CPU's are outselling Intel CPU's months running...

https://www.extremetech.com/computing/281741-new-retail-data-shows-amd-outselling-intel-21

https://www.techpowerup.com/254503/amd-outsells-intel-2-1-on-european-retailer-mindfactory-de?cp=4

And you think $11 is a huge price hike. Ha!

 

If AMD follow same paths as I is able to you should see 5000 CPU’s.


Not officially from Intel but I can probably see 10 cores on my old Z170.

6700/7700/8700/9900 and maybe coming 10 cores Will you be able to do similar on first gen Ryzen platform? Mod it? We have to see.

If some still buy the anniversary edition for +$500 I think owners which bought it new middle last year got an good value from it if they sell it and upgrade to 9900K

I talked about AMD prices mentioned in the Tomshardware link. What will be the end result we still don’t know.

 

MSI Issues Clarification on Next-Gen AMD CPU Support on 300-series Motherboards
PRESS RELEASE by btarunr Today, 10:52 Discuss (100 Comments)
https://www.techpowerup.com/254669/...en-amd-cpu-support-on-300-series-motherboards

"It has come to our attention that MSI Customer Support has regrettably misinformed an MSI customer with regards to potential support for next-gen AMD CPUs on the MSI X370 XPOWER GAMING TITANIUM motherboard. Through this statement we want clarify the current situation.

At this point, we are still performing extensive testing on our existing lineup of 300- and 400-series AM4 motherboards to verify potential compatibility for the next-gen AMD Ryzen CPUs.

To be clear: Our intention is to offer maximum compatibility for as many MSI products as possible. Towards the launch of the next-gen AMD CPUs, we will release a compatibility list of MSI AM4 motherboards.

Below is a full list of upcoming BIOS versions which include compatibility for the next-gen AMD APUs for our 300-Series and 400-Series AM4 motherboards based on the latest AMD Combo PI version 1.0.0.0. These BIOS versions are expected to be released in May this year.
"

 

New CPU Methodology: Best CPUs for Programming, Premiere, V-Ray, & More
Gamers Nexus
Published on Apr 16, 2019
Our new CPU workstation testing methodology is finally ready for unveil. We tested some of the best CPUs for programming (compile), Adobe Premiere, Photoshop, & more. Read the test methodology & article here: https://www.gamersnexus.net/guides/34...
Our test bench, software version numbers, and components can be found in the article above. This is our new CPU testing methodology for 2019 -- half of it, anyway -- and shows performance of Ryzen and Intel parts in production workloads. We also tested most CPUs with some level of overclock to establish the range of performance.

 

What you'll get...

Ryzen 7 2700X 50th Anniversary Edition Gets Laser Inscription With Lisa Su Signature
By Hilbert Hagedoorn on: 04/25/2019 07:57 AM

In an earlier post, we already shared that AMD is preparing for the festivities of their celebration 50th anniversary beginning with a Ryzen 2700x 50th-anniversary edition to be released on or around May 1st, 2019.

However, the packaging is updated to a new black and gold design. The proc itself will be a commemorative special edition as the CPU gets a laser-etched signature of AMD President and CEO, Lisa Su.

 

Speaking of what you'll get, did you see the most recent Intel leaked roadmap?
https://www.pcbuildersclub.com/en/2...-in-desktop-before-2022-and-maximum-10-cores/
https://segmentnext.com/2019/04/25/intel-cpu-roadmap-comet-lake-s/
https://semiaccurate.com/2019/04/25/leaked-roadmap-shows-intels-10nm-woes/
https://finance.yahoo.com/news/intel-10nm-desktop-dreams-may-124030746.html
No 10nm for you! LOL.

But in all seriousness, this is bad news for Intel!

 

Have seen it. And you're 100% sure this is the current and last (final) roundmap from Intel?

 

See:
http://forum.notebookreview.com/thr...-lake-cpus-z390.811225/page-149#post-10901721

 

I'm not an insider, so cannot guarantee anything. But, the limited 10nm low power dual core mobiles, then the replacement a year later, but no H or G actually does make sense. Also, Cascade and Cooper lake for HEDT and Server have long been known to be 14nm. Seems they are finally porting the Ice Lake/Sunny Cove back to 14nm.

With that said, and what @tilleroftheearth isn't saying, is 14nm is near its limits. The microarch will get better, meaning Comet Lake and beyond will have increases in single threaded performance due to improved IPC from architecture improvements. All that means is that Intel won't get the process improvements. So it does NOT mean Intel will not have more to get out of the tank, just means that the hopes of lower power is lost.

In fact, prior roadmaps showing Intel's release for the 9900K were only off by within two quarters on the products, depending which roadmap you looked at (the first released one turned out more accurate on that, whereas the second, later released one must have been an earlier version because it had the 9900K coming in Q1 of 2019, rather than late Q3 to Q4, like the first one).

If this one is anywhere near as accurate as those previously published, Intel has a process node problem.

 

Ryzen 7 2700X and Radeon VII Gold Editions Tomshardware.com | April 29, 2019

Despite the numerous leaks and rumors, the Ryzen 7 2700X 50th Anniversary Edition will not differ from the standard version in terms of clock speeds. There were high hopes that AMD would delight us with a cherry-picked chip with higher clocks, much like what Intel did with its Core i7-8086K processor.

Nonetheless, the Ryzen 7 2700X 50th Anniversary Edition will surely find its way into hardcore AMD aficionados' collections.

The Ryzen 7 2700X Gold Edition also includes a sticker with AMD CEO Lisa Su's autograph and a coupon for a free AMD 50th anniversary t-shirt. The processor is available today at numerous retailers.

 

AMD 7nm ZEN 2 64 Core & 32 Core CPU Specs & Benchmarks Leaked Wccftech.com | May 19, 2019

In fact it runs at a base frequency of 1.7 GHz and a turbo frequency of 2.4 GHz. Which is surprising considering it has half the number of cores. Although this particular chip could be a low power part, which would explain the low clock speeds. And unlike the 64 core ES, this part only appeared on the database once so far. Add to that fact that this entry is for a Dell PowerEdge R6515 server, this means that we’re bound to see more of AMD’s EPYC Rome parts as they’re sampled to the company’s partners.

 

Intel Loses 5X More Average Performance Than AMD From Mitigations: Report

https://www.tomshardware.com/news/intel-amd-mitigations-performance-impact,39381.html

 

An ES is by no means a final product.
So, its possible the clocks aren't final... or if they are, we don't know how this will affect performance (depending on IPC increases).

 

In my searching, I didn't see any leaks or rumors that said the performance of any of the AMD 50th Gold Anniversary CPU / GPU / Motherboards were anything other than a celebration of the event editions, and none advertised more performance or additional functional features.

No one advertised anything extra besides a Lisa Su "signed" element and / or T-shirt to go with the "different" accented (colored) features and box of the product.

There was also an AMD game pack change around the same time - change in selection of pack-in games - but it wasn't advertised as a Gold Edition special.

There were a couple of opportunistic reviewers that tried to make like there was going to be a difference to get clicks, but otherwise no one suggested or advertised or leaked that any performance boosts were included with the 50th anniversary components.

 

Agreed.
When the Gold Anniversary CPU's and gPU's from AMD were announced, I also hadn't come across any claims for clock increases or performance changes.
The only thing that was advertised was Lisa Su signature, and that's it.

 

Good thing we have one week until the keynote and should receive final clock speed around then.

Sent from my SM-G900P using Tapatalk

 

https://www.reddit.com/r/AMD_Stock/comments/bqu1se/has_zombieload_completely_defanged_intels_xeon/

Long_on_AMD 21 points 15 hours ago*
"Michael at Phoronix will be posting a new article later today on high core-count Xeons. These presumably cap out at 28 cores/56 threads, as the Cascade Lake-AP CPUs only go into dedicated Intel built systems which he is unlikely to have access to. Cascade Lake-AP isn't a real product offering to begin with; it's a panicked, smoking hot, water-cooled PR stunt to make Rome look less game-changing.

But yes, at the exact moment that AMD is about to release its truly game-changing Rome monster, Intel is actually delivering significant regression in performance, thanks to their decades of cutting corners in security. The schadenfreude is delectable! And next year, the game changes again, and in a big way, with Milan..."

A Look At The MDS Cost On Xeon, EPYC & Xeon Total Impact Of Affected CPU Vulnerabilities
Written by Michael Larabel in Software on 20 May 2019.
http://forum.notebookreview.com/thr...atches-and-more.812424/page-126#post-10913136

"...If looking at the geometric mean of all the benchmarks carried out, the EPYC 7601 averages out to about a 1% hit with its Spectre mitigations.

The dual Xeon Platinum 8280 Cascadelake setup with its mostly hardware-based mitigations was slower by 4% with the relevant mitigations enabled.

(l1tf: Not affected + mds: Not affected + meltdown: Not affected + spec_store_bypass: Mitigation of SSB disabled via prctl and seccomp + spectre_v1: Mitigation of __user pointer sanitization + spectre_v2: Mitigation of Enhanced IBRS IBPB: conditional RSB filling).

Meanwhile the dual Xeon Gold 6138 server that unfortunately doesn't have the hardware mitigations saw a 11% hit from the benchmarks run with these Spectre / Meltdown / L1TF / MDS mitigations or 15% if disabling Hyper Threading as an additional measure based on the benchmarks carried out today."

 

Spoiler: CB15 Score of a 16-core Zen 2 CPU clocked at 4.2GHz boost

Looks like 4287 was the score. My [email protected] gets 3652. That is a 17% increase, and there is a chance the 4.2GHz may or may not have been all core on that chip. For a comparison, the 7960X does not hit that score until 4.7GHz or 4.8GHz. This means when that 16-core part is released, it very well could eat the HEDT market's lunch!

Now, just want to know what lays ahead for Threadripper!

Also, that explains that the Epyc Rome chips upcoming will be able to hit around the same frequencies of the first gen Epyc chips.

6-days (actually, I'll be watching the presentation Sunday night due to where I live). Times they are a changing!