Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
Malloc Voidstar
May 7, 2007

Fuck the cowboys. Unf. Fuck em hard.

https://blogs.nvidia.com/blog/2017/...ed-data-center/
Intel, AMD Just Delivered Two Great Reasons to Upgrade Your Data Center

🤔

Adbot
ADBOT LOVES YOU

rex rabidorum vires
Mar 26, 2007



Risky Bisquick posted:

Which board? Figure that out and search the web cache for cpuz validations. This isn't ideal, but it's a start

[url]https://www.google.ca/?q=+3200+ryzen+site:http://valid.x86.fr/[/url]

I know I keep pimping this site, but it's the best amalgamation of ryzen and ram info I've been able to find. https://rymem.vraith.com/

SwissArmyDruid
Feb 14, 2014



The one time AMD actually sets prices that are almost *reasonable* compared to Intel.... and they gotta kill me by doing poo poo like maybe including AIOs with certain Threadripper SKUs.

http://www.guru3d.com/news-story/am...ng-bundled.html

Anarchist Mae
Nov 5, 2009

by Reene


Lipstick Apathy

SwissArmyDruid posted:

The one time AMD actually sets prices that are almost *reasonable* compared to Intel.... and they gotta kill me by doing poo poo like maybe including AIOs with certain Threadripper SKUs.

http://www.guru3d.com/news-story/am...ng-bundled.html

That... actually makes it a more attractive purchase to me. Assuming it's a decent AIO.

PerrineClostermann
Dec 15, 2012

by FactsAreUseless


But I want to build a custom loop!

Anarchist Mae
Nov 5, 2009

by Reene


Lipstick Apathy

PerrineClostermann posted:

But I want to build a custom loop!

Then build a custom loop?

But seriously, it'd mainly be attractive because it'd be a stock cooler. There would be no need to worry about getting the right fixtures to fit an already on the market AIO or pump to the CPU and you should get pretty good cooling.

And if you don't want it you could just not use it (or perhaps buy one without it maybe?).

Kazinsal
Dec 13, 2011






That is a hell of a lot better than the "stock cooler" for Intel high end chips being both a price-gouged Asetek AIO and a completely separate purchase from the chip itself.

Ihmemies
Oct 6, 2012



I wonder if the rad fits to the top 120mm slot on a silverstone raven 02: http://www.silverstonetek.com/raven...v02b-inside.jpg

If yes, I could see myself actually using a watercooled setup for the first time in my life. I don't want to buy an expensive $150+ case just to fit an AIO though.

eames
May 9, 2009



found a video comparison of 3200 Mhz CL14 vs 2133 Mhz RAM (R7 1700 @ 3.9 Ghz, 1080ti, 1080p)

https://www.youtube.com/watch?v=DlAboBQgAHk

Munkeymon
Aug 14, 2003

Motherfucker's got an
armor-piercing crowbar! Rigoddamndicu𝜆ous.





eames posted:

found a video comparison of 3200 Mhz CL14 vs 2133 Mhz RAM (R7 1700 @ 3.9 Ghz, 1080ti, 1080p)

https://www.youtube.com/watch?v=DlAboBQgAHk

Pretty stark difference in some of them. I'm guessing Witcher is very single-threaded?

Random aside but does anyone know what the game before Witcher was?

Mr Shiny Pants
Nov 12, 2012


Munkeymon posted:

Pretty stark difference in some of them. I'm guessing Witcher is very single-threaded?

Random aside but does anyone know what the game before Witcher was?

Crysis 3.

Paul MaudDib
May 2, 2006

"Tell me of your home world, Usul"


Munkeymon posted:

Pretty stark difference in some of them. I'm guessing Witcher is very single-threaded?

IMO it's more of a memory problem, there's very clear memory scaling even on Skylake/Kaby Lake.

On the whole W3 is actually pretty good about spreading across cores but there's no question that cities run noticeably worse, for everyone. Particularly Novigrad, the biggest city. I don't know if anyone's ever sat down and done the science to determine the exact behavior and mitigations (more RAM, more bandwidth, lower latency, etc).

EmpyreanFlux
Mar 1, 2013

The AUDACITY! The IMPUDENCE! The unabated NERVE!


So I was watching Hardware Unboxed (really good tech channel IMHO), and they did one on whether the Mesh clock was holding back Skylake-X performance. It's not, but in the comment section this came up

quote:

It's actually worse than Zen (I'll explain) L3 design, but Intel is making up for it with brute clocks & larger L2. L3 typically have more cycles (slower latency vs L2 access), on Zen, AMD makes up for this by making it both large and shared with the other cores in the CCX. It's not 2MB per core, it's 8MB per 4 cores, all shared with enough associations for parallel read/write cycles.

Zen's L3 acts as a core multiplier for cache hits, the more activity within that CCX, the higher the L3 hits will be because the byproduct cache data from each core gets shared with the others, any core that needs data, it checks that large L3 and hits it instead of tripping to RAM.

Intel's Skylake X L3 is a victim cache that is crippled because they did not make it a shared cache with the other cores directly, but any access has to hop over the nodes on their Mesh. It's core specific, smaller and non-inclusive, a really bad combo for an L3.

It's a compromised design focused for high core counts not meant to be great at gaming. They could have made it great for gaming too, but it would require more power and die size.

Sounds like Intel hosed up Skylake X L3 design and AMD might come away with the better overall scalable solution even without considering MCM. It's not Netburst bad but if Intel continues to push the mesh style uarchs which seem to regress performance they may put AMD in a position to pull another Athlon XP/64 again. Gaming desktop performance might not be a large market (or to say, it is, but it's a small market within context of the semiconductor industry) but it's a very loud, public section of the market, and one that is easy for a layman to get.

https://www.youtube.com/watch?v=mTQ6ymQIY64

PC LOAD LETTER
May 23, 2005
WTF?!

Slippery Tilde

FaustianQ posted:

Sounds like Intel hosed up Skylake X L3 design and AMD might come away with the better overall scalable solution even without considering MCM.
"hosed up" might be a tad harsh on Intel. IMO it'd be more reasonable to say Intel tuned their L3 better for server/HPC work loads at the expense of gaming performance and pooooossibly generic desktop applications too. Though to me, given some of the benches of Epyc vs competing Intel Xeons, it doesn't seem to be doing a whole lot of good for Intel in that market either. Maybe its one of those things that is supposed to matter more in the future with other part/future CPU's when they start getting really crazy with core counts and putting out 64C/128T chips and they just didn't expect Zen/IF bus to throw a curve ball at them this well??

Unimpressive gaming performance by SkylakeX definitely gives TR and 8C16T Ryzen in general an edge they wouldn't have had otherwise but probably not a decisive one in of itself. I think its when you combine that issue with some of the other X299 platform SNAFU's + higher pricing that Intel is demanding for Skylake X that TR and 8C16T Ryzen starts to look like the way to go. I'm sure some will still buy Intel no matter what though. After all even Netburst was a solid seller for Intel!

FaustianQ posted:

if Intel continues to push the mesh style uarchs which seem to regress performance
They're almost certainly going to be stuck with it for a while since re-engineering the mesh bus and/or cache to compensate is going to take time. If that is what is required it won't be as bad as all the crap AMD went through trying to get BD to perform well but it won't be something that gets fixed in 3-6 months either with a new stepping or BIOS update.

eames
May 9, 2009



Based on the benchmarks and videos I've seen so far the Intel Ringbus CPUs are their best chips for gaming and probably will be until at least Ice Lake. X299 isn't terrible but they shouldn't have marketed it for gaming.
The Intel CEO himself said something along the lines of "get a 18 core i9 for high resolution gaming". Truth is that the 7700K continues to run circles around these high core count, low frequency parts and Coffee Lake should do even better.
The 7920X just leaked today with +2 cores and -400 Mhz compared to the 7900X.

AMD is in an excellent position if they can squeeze more frequency/IPC out of Zen+.

EmpyreanFlux
Mar 1, 2013

The AUDACITY! The IMPUDENCE! The unabated NERVE!


I mean yeah it sounds pretty drat smart on the surface but as has been pointed out, it really doesn't seem to do anything for HPC either as EPYC is able to come in at better performance most of the time with equal or 20% better power draw, and is 50% the price or even 25% the price depending on dual or single socket configuration. It also means that for the foreseeable future, AMD is the better choice for HEDT (what AMD calls SHED, and I feel like there is a joke here) and that is not a sentence I'd thought I'd ever commit to writing.

PC LOAD LETTER
May 23, 2005
WTF?!

Slippery Tilde

eames posted:

AMD is in an excellent position if they can squeeze more frequency/IPC out of Zen+.
Oh yes absolutely. If Zen+ really does end up with ~15% better IPC + ~5Ghz clocks then they'll be able to meet or beat Intel on nearly any work load and for any market too.

At that point it'll come down to how many chips GF, or pooooooooooooooossssibly Samsung I guess too, can fab up for them.

AMD was always fab constrained in the past when they had a competitive architecture, I think for the longest time (even going back to the K6 era) the best they could supply was around 30% of the world's x86 demand (the Dresden fab was supposed to allow them to supply around 50%+ x86 demand initially but then multicore started to take off which blew up die sizes and that was that...), which guaranteed Intel to be able to still sell a heap of chips no matter what.

If after going fabless, and after all the poo poo they've been through, they end up being able to really meet a majority of x86 CPU demand for once in their history I'm gonna laugh and laugh and laugh.

Risky Bisquick
Jan 18, 2008

PLEASE LET ME WRITE YOUR VICTIM IMPACT STATEMENT SO I CAN FURTHER DEMONSTRATE THE CALAMITY THAT IS OUR JUSTICE SYSTEM.



Buglord

5Ghz? Anything around 4.5Ghz with Zen+ will close the casket on Intel's gaming dominance.

SwissArmyDruid
Feb 14, 2014



Ah, the 5 GHz chestnut again.

GRINDCORE MEGGIDO
Feb 28, 1985




I'd be happy if the fabric ran full speed asynchronously, it could run ddr4@3600 easily and it hit 4.2ghz stock easily at the same wattage as stock does now.

GRINDCORE MEGGIDO fucked around with this message at 15:12 on Jul 19, 2017

eames
May 9, 2009



Another Ryzen memory performance comparison.

http://translate.google.com/transla...n&langpair=auto

Their conclusion for gaming is that a 1600X with DDR4-3200 is faster on average than a 1800X with DDR4-2666 and clockspeed is more important than timings.

Cygni
Nov 12, 2005

raring to post



I dunno, Zen2 could blow everyones labia off, but this is Keller-less AMD on an internal Glofo node were talking about.

It's also not going to be on shelves until late 2018/2019, which means its theoretically gonna be facing 10nm Cannon Lake/Ice Lake from intel, a process which took them 3 years longer than they expected to develop. Hard to say what any of this stuff means yet.

NewFatMike
Jun 11, 2015



What is it, 2020 when AMD are done with the WSA? That'll be rad.

Seamonster
Apr 30, 2007

IMMER SIEGREICH


So is zen+ the die shrink? Or is that zen2?

SwissArmyDruid
Feb 14, 2014



Combat Pretzel
Jun 23, 2004

No, seriously... what kurds?!

14nm+ for H1 2018 or what?

Cygni
Nov 12, 2005

raring to post



Seamonster posted:

So is zen+ the die shrink? Or is that zen2?

Zen2 is what AMD is now calling what used to be Zen+ on their slides. The 14nm+ version of Zen doesn't have a specific code name that I've seen, and I don't think they are expecting too much of a performance boost.

EmpyreanFlux
Mar 1, 2013

The AUDACITY! The IMPUDENCE! The unabated NERVE!


Combat Pretzel posted:

14nm+ for H1 2018 or what?

Yea, Pinnacle Ridge apparently. My understanding is 14nm Plus is the LPP process with modifications requested by AMD, much like the 12nm process from TSMC for Nvidia. Considering their roadmap for 2018 has Polaris and Vega on it, I'm expecting yet another refresh for the GPUs on this process as well. All I can say is gl AMD.

EmpyreanFlux fucked around with this message at 00:23 on Jul 20, 2017

Combat Pretzel
Jun 23, 2004

No, seriously... what kurds?!

What sort of clock bump can be expect from LPP? 4GHz base on the top end CPUs would be nice. Or 3.7-3.8GHz on the top end Threadripper.

EmpyreanFlux
Mar 1, 2013

The AUDACITY! The IMPUDENCE! The unabated NERVE!


Combat Pretzel posted:

What sort of clock bump can be expect from LPP? 4GHz base on the top end CPUs would be nice. Or 3.7-3.8GHz on the top end Threadripper.

14nm LPP is a mobile focused process that doesn't seem to handle large die, high density design well so I don't expect better clocks but rather better perf/watt at current clocks. If you want 4Ghz base I'd be eyeing Zen2.

eames
May 9, 2009



If I remember correctly Jim Keller designed Zen and one improved version of it for AMD. I would hope that Zen+ is simply a die shrink and Zen 2 is that second design.

PC LOAD LETTER
May 23, 2005
WTF?!

Slippery Tilde

Zen+ and Zen2 are the same thing. Its a branding mix up for a unreleased product so its irritating but doesn't actually matter since almost no one but us and a few other industry types even looks at AMD presentation slides.

Anarchist Mae
Nov 5, 2009

by Reene


Lipstick Apathy

https://videocardz.com/newz/arctic-...en-threadripper



Holy poo poo lol.

GRINDCORE MEGGIDO
Feb 28, 1985




Hefty!

Scarecow
May 20, 2008

3200mhz RAM is literally the Devil. Literally.


Lipstick Apathy

I neeeeed more info about threadripper fuuuuuuuuuck

3peat
May 6, 2010



thiccripper

Truga
May 4, 2014




Lipstick Apathy

Cygni posted:

A theoretical 8 core TR would be 2/2+2/2. You could also do a comical 4 core with 1/1+1/1, which would join the Epyc 7251 8 core (1/1+1/1+1/1+1/1) in "this is for a very specific use case" territory, ha. Lots of cache/memory bandwidth, lots of PCIe, not a lot of raw proc power.

So, ARMA gaming rig?

isndl
May 2, 2012
I WON A CONTEST IN TG AND ALL I GOT WAS THIS CUSTOM TITLE

What's the average life expectancy of an ordinary vertically-mounted motherboard using that giant cooler?

Ihmemies
Oct 6, 2012



Maybe the mobos come with a full backplate too?

E: having AIO come default with TR CPU's might actually be a good idea

Adbot
ADBOT LOVES YOU

well why not
Feb 9, 2009





That plate is insane. They're pretty serious about this being a rack/server platform, I guess?

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply