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

If the scheduler runs more efficiently it may, or may not, translate into more FPS for a game. It may just turn out to be more efficient and thereby create less heat etc..

 

My point is I know it is an issue, in fact I was the first to point it out. It is just do not expect a magic pill here but there will be gains in at least some things.

Funny how close the 4+0 mirrors the 4+4.........

 

Not with the 500 series. With big Vega, it will be up there. Hopefully someone also tells them to officially support forced crossfire. Nvidia, although not officially supported, can force SLI if you know how. This is a big deal for some purchasers, which gives Nvidia an edge. Even if not the most efficient/elegant solution, it would be nice. The 580 is expected to be $200 like its predecessor. So, May will have Vega and the fun of comparing. Also, how many games support using Nvidia and AMD running together? I know it can be done with DX 12, but I don't know how many utilize it...

 

Now, now, why did you felt like bursting my bubble Yeah, I know, but I kinda hope that it would be a double release (500 + Vega), or at the very least there would be a teaser about what to expect from Vega. After all the Ti is in the wild for a while now and there's ZERO official information about anything competitive. That's why I think that they've pulled the R5 with a month or so. As I said earlier - the 1700 was supposed to steer some of the gamers (rather streamers, as the demo was clearly geared towards them, but thank internet for making it all about gamers) from 7700K, but obviously that didn't quite happen, or at least not as much as they hoped. That's of course pure speculation. In any case, I'm eager to see what the 6c and 4c have to show, seems that HWUnboxed thinks the same - "1800X for 170 bucks". Now let's see how much water that one holds.

That would be nice, but unlikely. From everything up until now it seems that they want to do most with the least money spent. It should come as no surprise, since they don't have a crap ton of money to start with. So, I guess eventual DX12, what was it called, unified (?) GPU is their preferred way to do it. It actually would be better, but we might see some of the performance drop that Ryzen sees currently (when shifting within CCXs). Now that they are back to talk with developers, let's hope that it would be worked on before it occurs (unlike Ryzen). EDIT: Imagine an XConnect x8 link that combines the internal RX 580 or even Vega (yeah right, but not unlikely if we use an SLi monster and its cooling for a single GPU), with another external one, bumping the performance 100% and using ALL of the display outs - both laptop and eGPU.

 

https://videocardz.com/67242/amd-vega-with-64-compute-units-spotted
http://wccftech.com/amd-rx-vega-benchmarks-8gb-hbm-2048-bit-bus/
http://segmentnext.com/2017/03/13/amd-vega-gpu-benchmarks/
https://www.pcgamesn.com/amd/amd-rx-vega-sisoft-benchmarks

So, when it comes to benchmarks, the majority focused on the early Vega variant (not the C3 newer variant), yet all variants are clocked lower than the estimated 1550MHz to achieve the stated 12.5TF. Further, it is likely that the compared hardware is not running stock. In other words, it may not be an average representative of what performance people get from Nvidia (much how overclocked, sometimes significantly, CPUs from Intel were compared to stock Ryzen chips before release. So, this information leaked is a teaser, but is often shaded. If you look at the clocks, the speed of the card will be increased by about 25% over what is seen in those benchmarks. They say the 1080 Ti is 35% over the early C1 variant. That means it would need 10% increase through driver optimization to match the Ti, as well as any changes in later revisions, VBIOS optimizations, etc. So, mathematically, it suggests it is going to be closer than these articles make out. Considering with AMD, the performance maximization comes 3 months after release through drivers (usually significant, amounting to 10+%).

I think the separation between CPU and Vega is to allow people a month of paychecks to buy the platform first, then the flagship GPU second. You have to think of the consumer saving $800 in a month or two, or having to do $1400-1600 for the video card and platform. Also, in the past, it took a hardware mod to unlock cores (phenom). It may exist in the microcode, but until a person comes up with the work around, you won't be able to easily unlock it.

Edit: Above assumes somewhat linear scaling with increased speed.

I hear ya, but a man has to have dreams. I popped your dream so you popped mine. I guess we are equal now! But, they pushed using two 480s to be stronger than a 1080. If they supported forced, unoptimized, crossfire, then this becomes a bigger talking point moving forward. It also could give more incentive for companies to want to officially support it so that people wouldn't complain about multi-GPU performance compared to optimized partners.

 

Here is a little information on what I have run so far with my new rig.

4191.5 Validation before I got my 1080TI and was using a cheapo video card to set my system up.

http://valid.x86.fr/vzmd3w

BF1 at 3.8 @ 1.25 Voltage with a 1080TI overclocked +150/+300 with a 2560X1440 - 144hz display.



Cinebench at 4.025 @ 1.4 Volts in the Bios.



BF1 at 4.0 @ 1.3625 Voltage with a 1080TI overclocked +150/+300 with a 2560X1440 - 144hz display.



Aida64 for 9 hours at 4.025 and 1.3625 in the bios. Not sure that Aida64 is optimized for Ryzen yet though.

 

Where'd you get that chip? Which MB (the MSI carbon I take it)?

Edit: You should also mark it as 4.025. 4.25 is 4250, which at 1.4V would put it at potentially the best chip I've seen out there...

 

Got the chip from amazon it is a china chip without the D marking. Motherboard is MSI Carbon from microcenter. Not many choices on boards, I would have went with the Asrock Fatal1ty or Taichi but is hard to get ahold of them, was lucky enough just to get the carbon. Doesn't matter much though as the board does not get hot at max overclock. I will pick up some better (faster) ram once more testing has been done and the bios updated.

Edit: lol, ya your right that would have been a big lottery winner at 4.25, made the change.

 

I hear you. Having to RMA my chiller. Also, because of the MB shortage (and hoping they use this time to get a better board ready) I'm holding off until at least May...

 

I usually agrees with you man but I have to disagree here.

There are people like me who refuses to touch the multigpu side of things. It is messy and takes work to get it fixed. It also usually happens months after release and some games never support it due to engine limitations etc. It is hard enough to find a good game as it is nowadays, I dont want to having to pick games based on my GPU.

 

I said official. I like to speculate and calculate just like the next guy, but it is better to have official information in the end of the day. After all, too much "leaked" information can skew expectations. It happened with Polaris (I was there as well) and again with Ryzen (tried to hold off) and I'm pretty certain that there would be disappointed people with Vega as well. If you return and go by what AMD has teased, no one would've been disappointed neither from Polaris (aside from RX 480's power issues), nor from Ryzen. So yeah, I'm talking about official AMD tease, or leak if you want, not what we get from all over the web. I REALLY hope that would happen sooner, than later. People are buying Tis left and right.

Again, my guess is to show-off what unified GPU is capable

Same.

 

Ryzen 7 1800X: Windows 7 vs. Windows 10 Gaming Performance

 

So, there is no real world, tangible, performance gains on Win7, correct? When 'gaming' is the workload?

And at least a couple of (just as big) Win10 'wins' when the same workload is used.

The cost (as always) of using the decade old O/S today is that no further improvements will ever be introduced, whereas Win10 going forward is continually improved (possibly as early as April 11 when the 'creator's update' is released).

All the video's posted so far show the same thing; Ryzen is good enough for anyone with a smaller budget.

Hope to see Ryzen II compete with true 'worlds' best processors available at any $$$$.

 

So Win 10 NSA edition is approaching the equivalent of SP2 release on the never ending beta turd. Good to know.

Also, as to gaming, some do better on 7, some on 10. It depends on game OS optimizations. Meanwhile, there is tons of productivity software that either has not moved to 10, or has not optimized to provide the benefit. I do agree, in time, 10 will be optimized enough that the change over will occur (nature of the beast).

Not going to entertain the hardware discussion.

Edit: this means I'll need to get a new image and spend a week removing components and testing for stability and compatibility. It sucks cause I did that a month or two ago for the AE (was on the LTSB for Ent.). Looking at past releases, by a couple months after the second or third SP (thinking Vista and XP, SP1 for 7), things started looking better. But testing for deployment SUCKS!

 

Just in the process of reading this article: https://www.hardocp.com/article/2017/03/15/amd_ryzen_5_processor_family_introduction

With regard to the upcoming Ryzen 5 CPUs and CCX layout, there is one excerpt in there that says: "In closing, while AMD would not answer questions on this previously, Anandtech says these CPUs will be 3x3 and 2x2 when it comes to CCX derivation".

So, it looks like the we won't ever be getting away from the latency between CCX when it comes to the 4 core versions (ie. will not be just one CCX), so from a gaming perspective we can't really expect the 4 core CPUs to be any better than the Ryzen 7 that has already launched (with the exception that maybe we'll be able to overclock an additional 100-200Mhz due to the simplicity of 4 cores - my guess). Looks like any future improvement to gaming performance in Ryzen is gonna have to come on the software side of things (Windows / Game Developers).

 

If you want to avoid latency issues on the 8 core you can just lock the game process onto one CCX right now. Beside price the 4 core can't bring much.

BTW, is it fair to claim that Ryzen is the first time a consumer CPU has clearly demonstrated the need for NUMA optimization running many typical consumer software?

 

There is one problem with Numa, although it does keep it to separate CCXs by treating them as different CPUs, you can run into memory management issues, among others. This was talked about in that article discussing the scheduler awhile back (the one where they did the core ping test). Now, it may help with specific tasks as a stop gap, but raises its own issues. Intel had to use NUMA for its multi-core chips at one point, but getting scheduler latency awareness and designing software to distribute in different size chunks (as well as assigning tasks and processing dependent on other cores completion to the same CCX) will allow for a better implementation than NUMA (at least from what I've read, I'm not an expert with NUMA and multi-CPU systems)...

Edit: to be clear, the issue, if I remember correctly, is with the single imc instead of two. In the Naples server processors, each CCX has its own icm (which is why it has 8 channels). So on the server chips, no problem. Also, software isn't optimized for NUMA except in server applications and some workstations.

 

I see it much differently

Windows 7 performs as well as Windows 10 within the margin of error in measurement in all of the gaming tests, trading top performance measured pretty evenly across the games tested.

So there has been no OS improvement for general performance with DX11 games since Windows 7.

So there is no reason to upgrade to Windows 10 as tested on Ryzen.

If MS blocks updates on Windows 7/8.X someone will find a way around it, and important updates - are there any anymore?? - can be applied eventually.

Spoiler

Microsoft Drops Windows 7/8.1 Support for AMD Ryzen & Intel Kaby Lake CPU & Lies About It - BUSTED!
Published on Mar 16, 2017
Today Microsoft officially stated that it won't allow Windows Updates on Windows 7 or Windows 8.1 if you have an AMD Ryzen or Intel Kaby Lake or newer CPU. They claim the dropped support is for stability and reliability but I cut through the crap as an ex-Microsoft employee and tell you why they are really doing it. Comes down to Windows 10 market share and spying, plain and simple.

Raton Colorado 5 hours ago
I dropped 10 on a few systems both new store bought and custom cause it kept dropping boot or some other errors that kept the system from coming up, and now am on 7 again with no problems what so ever.

Windows 10 is still unnecessary, even with modern CPU's, and can be completely avoided.

 

https://arstechnica.com/gadgets/201...s-ryzen-at-games-but-how-much-does-it-matter/

Very good, indepth read on whether if it matters that Intel still beats Ryzen at gaming.

Some interesting debate going on in the comment section (Ars' comments rival the discussion in this forum thread)

From chaboud
"As someone who spent 5 years as the optimization point-person for both AMD and Intel at a division of Sony, I can say, without a doubt, that optimization *absolutely* is done for specific instructions set extensions, architectures, and processors. Some instructions are implemented very differently behind the scenes, and tight loops that leverage micro-architectural optimizations will definitely make vendor/family/revision specific calls to tightly optimized functions. This can make a very significant difference in the performance of tight loops (e.g. 2-8x running time), and the vendors of these processors have teams dedicated to aiding with those changes.

Similarly, if you hang around GDC, you'll find game and engine developers making similar modifications.

Sorry, Peter, but your assertion that most people don't understand optimization, while true, doesn't extend to the idea that all optimization is for all processors. In tight loops, very specific code can be, and often is, written."

Response from Peter Bright - Ars Story Writer

"I couldn't find any game developers willing to attest either on or off the record to this, and several saying the opposite. C++ and intrinsics appear to be the common limit, with assembler being exceptional. Stuff like emit to use instructions not supported by the compiler? VC++ is too common for that to be common. For 64-bit software, even inline assembler is not an option (mainly for stubbornness reasons I'm sure).

There's not a ton of widely available source code in this space to verify these things independently, but a cursory look through Unreal Engine 4, for example, suggests no assembler in the core engine (though some is used in some of the dependencies on some platforms), and nothing processor model specific at all, only instruction set specific. Of course, Unreal Engine 4 is, by design, pretty portable, so that probably limits the appeal of such things, but it's pretty mainstream nonetheless.

CRYENGINE V similarly appears to use only a tiny amount of non-C++, and much of that appears to be to called rdtsc or figure out of a chip has cpuid (so why they're using inline assembler and not intrinsics is anyone's guess). There is one matrix multiply routine using SSE optimized for the Pentium III that's in assembler but I don't think it's even used today. There is intrinsics, though again it appears to be keyed to instruction sets, not architectures."

 

Is this the middle finger to MS, as we all hoped it would be? Or am I just late to the party?

http://support.amd.com/en-us/kb-articles/Pages/am4-chipset-driver.aspx

 

Whoa!!

Spoiler: Screenshot of that Windows 10/7 64-bit driver text...

What about Windows 8.x?

And yes, I did download it for future use

 

You mean finger number three? The lengthy? I'm pretty Competent. I can count to three from left to right. And the opposite. Always come to the same finger. The lengthy. It must be the one you are talking about?

 

A nice coverage of all the little bits discovered, with gaming benchmark comparisons to back them up. Amazing how a slight delay in doing a review can help focus the testing and results

Ryzen 7 1800X Review! What's The REAL Deal With Gaming?

 

This Explains How Much Power Ryzen Really Has
https://www.reddit.com/r/Amd/comments/6021w3/this_explains_how_much_power_ryzen_really_has/

"I was playing battlefield 1 and i spaced out that i had a project for school due monday.

I was so accustomed to waiting after playing games to do my render work for projects that it did not occur to me to try to do it and play at the same time.

Well i tried and no stutter and 140 FPS avg in handbrake im at awe right now. AMD im glad i gave you my money."

" Portbragger2 13 points 17 hours ago

Loool ! Isn't it ironic.

I can play any game faster than on two CCX when setting affinity of my games to 1 CCX only. And at the same time I can do my compile jobs on the second CCX...

LOL LOL Ryzen is so ****ing amazing"

 

It actually isn't. Because of the lack of motherboards out there and all the current issues (many of which have been disproved or the culprit being slightly different than expected and the CPU locking down ram), AMD and Motherboard manufacturers put out the Win 7 drivers. It was a stop-gap needed to entice because the shortage changed purchase decisions.

Meanwhile, the new Creator's edition and a recent update are meant to try to move people to win 10. The creator's edition releases the same day as Ryzen 5. The means it was planned. The update this month blocks all future updates if Ryzen or Kaby are discovered.

So it isn't a middle finger, but a way to stop some bleeding from the FU of hardware partners.

Edit:
@Mr.Koala - "With that, it is possible for Windows to logically split the CCX modules via the Non-Uniform Memory Access (NUMA), but that would force everything not specifically coded to span NUMA nodes (all games, some media encoders, etc) to use only half of Ryzen." https://www.pcper.com/reviews/Processors/AMD-Ryzen-and-Windows-10-Scheduler-No-Silver-Bullet

 

Are you for real? Win 10 Creator's update has been planned for months beforehand.

See:
http://www.windowscentral.com/microsoft-announces-next-major-update-windows-10-creators-update

Posted October 26, 2016

The creators update isn't meant to try to move people to Win10 - although it undoubtedly will convince some to do so - it is just another step in the continuous process of updating the O/S to handle the hardware, software and interfaces that keep developing continuously too.

AMD chose April 11 for whatever reason to release Ryzen 5 on... has nothing to do with MS, sorry.

 

Yeah; we're reading (I guess) the same things and concluding differently; agreed.

Up till now, Win7 was the go to O/S for maximum gaming performance - actual testing doesn't bear this conclusion out. This gap will only increase with time as Win7 is frozen in time while Win10 keeps current with current tech and is always looking forward to work efficiently with the tech coming down the line.

By your thinking; Ryzen plus Win7 is the way forward (even for such a narrowly focused workload as gaming) - when all objective tests and reasoning shows that Ryzen hasn't matched Intel and MS will be dropping Win7 as soon as it's contractual obligations are met.

See:
http://www.pcmag.com/article2/0,2817,2475079,00.asp

Note the date: January 13, 2015 - O/S is 'frozen' in time. January 14, 2020 ends Security Updates. Good luck after that.

Sure; MS can extend the support (and it most likely will) - but that is hardly a basis to stay with a decade old platform on 2017 hardware and newer...

As to the ex employee spilling company secrets? Yawn. Business as usual... we are having an adult conversation here, right?

 

Are you ****ing daft? What you showed is them working on the release since last year, not the date of release. Software and hardware companies regularly coordinate to drop on the same day. This is to promote moving to the new software while upgrading your hardware. Further, this explains the update to block future updates to Windows 7 and 8/.1 if Kaby or Ryzen is present. By doing this weeks in advance, it allows purchasers to plan the additional cost of the OS into the purchase decision. It also gives the impression the new hardware will work better with the new software. Considering the creator's edition was not ready for Kaby and Ryzen 7, but that the HEDT Intel platform was pushed to August, having a coordination for Ryzen 5, containing 6-core CPUs expected to sell well and MBs expected to be in stock, makes sense.

Now, the launch dates were announced separately, but this does not mean that no coordination existed. So, this shows a lack of knowledge on corporate strategies, once again showing you have no clue on what you are discussing.

Sent from my SM-G900P using Tapatalk

 

So in one breath you accuse non-cooperation with AMD with most everyone else - including MS - and in the same breath you state there is cooperation/coordination. Cool.

Can't have a reasonable conversation like that. Circular reasoning is out of control in this thread.

Not daft at all. And please stop with the swearing, even the cool kids are getting bored with it. It certainly isn't helping your arguments any.

 

You two definitely need to have a tech show. I would watch it for the tension filled debates. Think about it, money would flow.

 

WTF are you talking about? Companies still try to work together to increase sales. That doesn't mean they didn't give an FU to the company in another way.

To explain, releasing on the same day takes practically nothing but a decision on the date. Giving an impression of working together well doesn't mean it does. M$ is playing off the hopes people have that the new version will work better and have optimizations for Ryzen as it is about a month and a half after the release of Ryzen 7 and new means those optimizations can be included. Implicit in this is that the optimizations WERE NOT INCLUDED FOR THE RYZEN 7 LAUNCH, a lack of communication between the two companies or an FU from M$, however you or any other readers see it.

As to the MB availability, the same applies for having a month and a half to shore up supplies for expected sales. Although a business decision to avoid aging inventory, it was an FU to AMD, but was grounded in historical sales data and then sucking Intel's dick on the 200 chipset launch.

So nothing I've said has contradicted other statements. You evidently cannot understand strategy and complexities of corporations. There is nothing circular, you just lack depth and comprehension.

Sent from my SM-G900P using Tapatalk

Edit: also, about a couple weeks after this, the microcode unlocking IMC and options on higher speed memory drops, which MB companies will have less leeway blaming AMD if not incorporated in short order.

 

I know what I'm talking about, but it doesn't help to try to explain it to you; you've already got your mind made up; you're always right.

How far down the circular reasoning hole have you gone?

Try to read your quoted text below and pretend that I wrote it. Are you laughing right now?

Here is a more truthful representation (or at least a more logical one):

Companies combine their strengths and hope/expect to each be better off than before (and especially better off than without working together). This isn't done through friendships; these are contracts that have multiple lawyers on both sides ensuring that their clients interests aren't only protected; but ensured.

A 'release date' is far from a simple decision from corporate to make. The details of that release matter immensely (as I previously criticized AMD beforehand of it's rushed release of Ryzen 7).

The reason that optimizations were not included for Ryzen 7 is implicit in the statement above.

For MB makers; if I was one of them now, I would be rushing to have as much stock as realistically possible (for their market share/etc.), but not more. Having an oversupply of product wont' make them more $$$$$$$ - selling the product that they do have at the highest possible price will. Econ 101.



Now, I'm really hoping you don't respond back with your usual wall of pseudo-'facts' to the above.

Let's continue a real conversation. I'd welcome it.

 

Yes, let's remove context from statements and then call someone else circular in reason. That helps clarify to prove a point!

Do you even know what circular reasoning is? It is beginning with the postulate attempted to be proven. My nuanced argument explaining the strategy of business, part of which includes not liking a hardware partner, yet still working with them because money is the end goal, is neither circular nor in contradiction. That is a problem we have, you assume innocence of corporate decisions, I do not. Why don't I? Because I'm the asshole they go to in order to make their plans work and comply with the law! Don't get me wrong, regardless of how I feel about it, personally, if you ask me to do something professionally and pay me for those services, I do them to the best of my ability. That is why I don't put myself in the place of having to do personally morally objectionable tasks. Corporations are amoral, but they are a dog after a bone. This is why leashes (regulations and laws) are a necessity, to prevent them from attacking and biting people or ****ting wherever they want (such as using retirement plans as equity in negotiations of mergers, allowing the purchaser to sell off that value and screw people out of pensions or for environmental dumping). Since I have to advise on actions, I have an idea of how management at these corporations think, as well as what spin would be involved. This is why I give zero benefit of the doubt, yet I do often take the time to mention business decisions versus other decisions that, although the company made, were just bad decisions. I have dropped those caveats regularly and [fairly] consistently.

Sent from my SM-G900P using Tapatalk

 

@tilleroftheearth and @ajc9988

I was quite seriously about a show, a podcast of sort. I'd definitely help fund it. You two obliviously have the intellect and the headstrong mindsets that would produce an award winning web series. The best shows tend to have two, sometimes three individuals that have a small amount of distain for each other but can articulate their points. I believe that's why the mods are letting you two go at it, because both of you make valid points. I say keep it going guys!

 

The experiment with limiting the 1800X to 6c/12t is very interesting. If the X1600 is the 1800X cut down, then that will make the 1800X a complete nonstarter outside of those that have already purchased it. It is a good thing I didn't checkout on Newegg for the 1800X and I'm going to wait until April 11/12.

 

The mods are now saying enough is quite enough. this is becoming nothing more than a fight than a thread that is on topic.

 

He does a really nice job on his review. Its really nice to see AMD back in the game again. I know I could not care less about 10-20 frames per second, as it does everything else so well.
I have been contemplating a Ryzen build, and reviews such as this keep pushing me toward that.

 

ajc9988, this is your first post that makes sense to me. It helps when you've clarified your background (lawyer) a little too.

I love playing devils advocate too (just like you seem to too) - but the 'truth' is black and white, not shades of grey (or shades of facts - from pseudo-facts, educated guesses or actual (gasp) real facts).

All cases aren't winnable. No matter how good you can argue.


Rage Set, thanks for your kind words and your generous offer, but I politely decline. My interests and time commitments lie elsewhere.

TANware, no problem, I can stop. I've made my point and others can now judge for themselves too on the topic I'm replying to. I can sincerely tell you that I'm not 'fighting' and just responding on topic; to the 'fact' that Ryzen is not the world's fastest processor (period) as is still stated on the OP and defended needlessly and endlessly when objective points (and BM's, sigh...) show otherwise.

Worlds fastest 'anything' titles don't come with caveats (such as bang for the buck) and mudslinging at competitors. At least not in my business scope.

Done.





P.S. ALL: Don't take this as me feeling I'm 'right'. I feel that anytime true discussion is halted it is a sad moment for all (just when it seemed it was going to really start). Thanks.

 

Prime95 for Ryzen released!
https://www.reddit.com/r/Amd/comments/608lis/prime95_for_ryzen_released_link_fixed/

"Prime95 version 29.1 build 13(!) is available. I've upped the major version number as version 29 signifies AVX-512 support. However, I've only upgraded trial factoring for AVX-512. FFTs will follow (much later).

Who might consider upgrading? Users that do trial factoring with prime95. Users that have had difficulty getting prime95 to assign affinities properly. AMD Ryzen users."
http://www.mersenneforum.org/showthread.php?t=22141

 

Does AMD have a Quad Channel 16 Core / 32 Thread CPU in the works? Guru3d.com

 

Hmmm. AMD's $1,000 flagship CPU?

Question for all. Would you pay $1,000 if this processor does surface?

 

I do not know what the price will be or performance, but Intel is not alone anymore

 

http://www.tweaktown.com/news/56757/amd-ryzen-cpu-16c-32t-coming-teased-999/index.html

Also, @Raiderman, this isn't the flagship, it is half the size of the flagship!

Edit: this means the 32C/64T may cost around $2000-2400, which trade blows with Intel's flagship Xeon.

Edit 2: so the Xeon E5 2698 v4 costs $3208.99 for comparison.
https://www.superbiiz.com/detail.ph...C7DKUSd219ixWzCzLIzFEMg69phNHetwDkaAsrd8P8HAQ

Sent from my SM-G900P using Tapatalk

 

Yes, but it is a Ryzen not a server cpu. This would be an enthusiasts chip.

Sent from my SM-G935T using Tapatalk

 

If they are not aggressively locked and MB/cooling solutions that can maintain ~4.0GHz on all cores exist, absolutely. ~400W for daily use is completely feasible with a good water cooling setup, the chip just need to not limit itself.

If not, hum... I have some very cheap Xeons.

 

It is LGA! That means different socket. The only other socket planned is for their server chip. So, this is from their server lineup. That makes sense as you need a board redesign to give you higher than dual channel. They may compare it to the HEDT lineup, but this is a server chip!

Edit: @Mr.Koala, I agree. It will come down to whether this can hit higher frequencies and if able to overclock.

@Raiderman - all server chips are built on Ryzen architecture, so calling it Ryzen is proper.

Sent from my SM-G900P using Tapatalk

 

Ahhh, okay

Sent from my SM-G935T using Tapatalk

 

Can I test it first? (Yeah; my workloads/workflows).

At the expected level of performance and the indicated $$$... I wouldn't need to wait for Intel's 'answer'...

 

The CPU arch is called "Zen". The consumer desktop lineup is "Ryzen".