AMD - Zen chitchat

[Xlucine]

I think Intel is screwed look at the R5 1400 pricing.

Also some more investigation of gaming performance and SMT in gaming:

http://www.tomshardware.com/reviews/...ming,4977.html
https://www.purepc.pl/procesory/amd_...wylaczonym_smt

More news on the Ryzen APU and Zen 2:

https://videocardz.com/67362/amd-pin...o-8-zen2-cores

I'm still convinced that bristol ridge won't hit retail as components for consumers, and a 4 core zen APU in a laptop sounds like a hoot - much better than the intel i3's pretending to be i7's!

[bsodmike]

Wraith Spire on the R5 1500X and R5 1600.

CAT - do you know details on how the XFR/boost is applied to these new SKUs with regards to the XFR/boost applied across all cores or just one core?

I just read somewhere else that the 1600X's XFR/boost is on a single core only. Thanks!

Cross that bridge when you get to it......future-proofing a gaming machine is not worth it.

Agreed - I tried doing this, and meh, it may workout for me or not. That said, I don't really upgrade all that often. My recent 6850K/X99/GTX1080 build is more than plenty, potentially for a good while.

I was also thinking of doing GTX 1080 SLI, but decided against it given how things run quite well in the current config.

True enough, but since I said, or at least heavily implied, that it's not a gaming machine (though it's also used for gaming) I much prefer that all bridges that can be pre-crossed are pre-crossed on the day I build it.

Just don't go overboard like I did; not sure if it justifies the cost in the long run

[shaithis]

I was also thinking of doing GTX 1080 SLI, but decided against it given how things run quite well in the current config.

The transition to DX12 has pretty much killed SLI/CF and very few DX12 devs have implemented mGPU. We are in a partial-limbo state with regards to multiple GPUs......which sucks when we are also in a transition to 4K and VR!

[CAT-THE-FIFTH]

AT thinks that the R5 1500X has 16MB of cache and the R5 1400 has 8MB of cache:

http://www.anandtech.com/show/11202/...n-5-april-11th

We have confirmation from AMD that there are no silly games going to be played with Ryzen 5. The six-core parts will be a strict 3+3 combination, while the four-core parts will use 2+2. This will be true across all CPUs, ensuring a consistent performance throughout.

So it seems the R5 will be a dual CCX design.

[CAT-THE-FIFTH]

https://www.reddit.com/r/Amd/comment...7_and_windows/

When you try to scan or download updates through Windows Update, you receive the following error message:

Unsupported Hardware

Your PC uses a processor that isn’t supported on this version of Windows and you won’t receive updates. Additionally, you may see an error message on the Windows Update window that resembles the following: Windows could not search for new updates An error occurred while checking for new updates for your computer. Error(s) found: Code 80240037 Windows Update encountered an unknown error.

Cause

This error occurs because new processor generations require the latest Windows version for support. For example, Windows 10 is the only Windows version that is supported on the following processor generations:

Intel seventh (7th)-generation processors
AMD “Bristol Ridge”
Qualcomm “8996"
Because of how this support policy is implemented, Windows 8.1 and Windows 7 devices that have a seventh generation or a later generation processor may no longer be able to scan or download updates through Windows Update or Microsoft Update.

Resolution

We recommend that you upgrade Windows 8.1-based and Window 7-based computers to Windows 10 if those computers have a processor that is from any of the following generations:

Intel seventh (7th)-generation "Intel Core" processor or a later generation
AMD seventh (7th)-generation (“Bristol Ridge") processor or a later generation
Qualcomm “8996" processor or a later generation
I understand how Windows 7 would not be updated going forward, but Windows 8? Really?? It has standard support from Microsoft until January 2018.

I don't know about anyone else here, but I love Windows 8.1 with Classic Shell. It has everything I need out of a Windows OS, however when it comes to Windows 10 it just feels wrong in some way to me.

Personally I'm still going to be running Windows 8.1 even when I get my Ryzen rig. Thanks Microsoft for another XP-era!

Anyone else have any thoughts on this?

Looks,like MS is now blocking Win7 updates if you install Ryzen or Kaby Lake.

[bsodmike]

The transition to DX12 has pretty much killed SLI/CF and very few DX12 devs have implemented mGPU. We are in a partial-limbo state with regards to multiple GPUs......which sucks when we are also in a transition to 4K and VR!

Yeah - fancy tech is really worthless, when it ultimately boils down to whether devs actually leverage such features.

Given that my X34 display is only 3,440x1,440 (QHD) and gets around 80fps maxed out in Rise of the Tomb Raider with the 6850K/GTX1080 combo, any gaming related future upgrade would be made to the GFX/display, to potentially move to 4K.

Honestly, I don't think I'd bother doing that though until 4K is pretty-much mainstream with 60-70+ fps capable systems.

FYI I still watch most of my home-theatre content in 1080p on a 42" Sony Bravia from 2010 (LOL!)

[kalniel]

AT thinks that the R5 1500X has 16MB of cache and the R5 1400 has 8MB of cache:

http://www.anandtech.com/show/11202/...n-5-april-11th



So it seems the R5 will be a dual CCX design.

Though it seems like there's no problem with Windows 10 scheduler understanding how to manage this fine:
https://www.pcper.com/reviews/Proces...-Silver-Bullet

[kompukare]

See that Hardware.fr also reviewed the 1700 and 1700X yesterday.
And they made these nice little volts, clock and watt charts:
http://www.hardware.fr/articles/939-...-pratique.html (Google Translate)


Obviously only their samples but interesting anyhow. Seems their 1800X is the best after all as it's able to hit 3.9GHz at 1.275V. The 1.200V data point is the most interesting as not only is the 1800X able to hit 3.8GHz with that rather than 3.7GHz, it does so using the least power too.

Never looked into this, but how much variation would there be at the same clock and voltage? That is taking a different CPU , running it at a set clock and feed it the exact same voltage. Would the wattage be very similar. That is, is the main variation in the silicon lottery mostly the max clock at a certain voltage, but that when clock and voltage are similar the wattage is too?

[kompukare]

Also see that PCGH is currently testing the new BF1 DLC.
Very much still a work-in-progress but excellent showing from Ryzen especially considering the Intel HEDT systems have 4x8GB and everyone else gets 2x8GB.

[bsodmike]

See that Hardware.fr also reviewed the 1700 and 1700X yesterday.
And they made these nice little volts, clock and watt charts:
http://www.hardware.fr/articles/939-...-pratique.html (Google Translate)


Obviously only their samples but interesting anyhow. Seems their 1800X is the best after all as it's able to hit 3.9GHz at 1.275V. The 1.200V data point is the most interesting as not only is the 1800X able to hit 3.8GHz with that rather than 3.7GHz, it does so using the least power too.

Never looked into this, but how much variation would there be at the same clock and voltage? That is taking a different CPU , running it at a set clock and feed it the exact same voltage. Would the wattage be very similar. That is, is the main variation in the silicon lottery mostly the max clock at a certain voltage, but that when clock and voltage are similar the wattage is too?

Also note that 1800X & 1700X temps are inflated by 20degC.
https://community.amd.com/community/...e?sf62109582=1

[kompukare]

So it seems the R5 will be a dual CCX design.

Well the 6C/12T were always going to be, but for the quads this does not bode well for performance as the PCGH tests showed 4+0 to consistently faster than 2+2 for the four games they tested except in a few places in Watch Dog 2.
http://www.pcgameshardware.de/Ryzen-...eview-1222033/

[CAT-THE-FIFTH]

The problem is I am not sure if AMD PR has again contradicted themselves - if the R5 lower end SKUs have 8MB,does that mean they can partially disable the L3 cache per CCX now??

[CAT-THE-FIFTH]

Also see that PCGH is currently testing the new BF1 DLC.
Very much still a work-in-progress but excellent showing from Ryzen especially considering the Intel HEDT systems have 4x8GB and everyone else gets 2x8GB.

The FX6350 is still holding in there at 60+ FPS though!!

[kompukare]

Duplicate post, deleted

[scaryjim]

I stole this quote from the Hexus news thread on Ryzen 5, but my comments are more relevant here...

AT thinks that the R5 1500X has 16MB of cache and the R5 1400 has 8MB of cache:

http://www.anandtech.com/show/11202/...n-5-april-11th...

I think there's a very important half-sentence in that article:

the default Windows scheduler knows how to manage that boundary as demonstrated by Allyn at PCPerspective earlier this week.

The link in the quote does a pretty good job of showing that not only does the Windows scheduler attempt to avoid assigning two loaded threads to the same physical core, but that it will also attempt to keep threads from the same process running on the same CCX. In other words, in cases where either SMT or inter-thread communication across a CCX are causing issues, it's likely to be down to the game code attempting to run an inappropriately optimised codepath. If they just let the Windows scheduler do its thing it'll do the right thing....

One other tidbit from the linked testing though: Intel's inter-core communication takes twice as long as Zen (although only half as long as Zen across the CCX boundary). There's some really good bits in Zen.

[CAT-THE-FIFTH]

I stole this quote from the Hexus news thread on Ryzen 5, but my comments are more relevant here...



I think there's a very important half-sentence in that article:



The link in the quote does a pretty good job of showing that not only does the Windows scheduler attempt to avoid assigning two loaded threads to the same physical core, but that it will also attempt to keep threads from the same process running on the same CCX. In other words, in cases where either SMT or inter-thread communication across a CCX are causing issues, it's likely to be down to the game code attempting to run an inappropriately optimised codepath. If they just let the Windows scheduler do its thing it'll do the right thing....

One other tidbit from the linked testing though: Intel's inter-core communication takes twice as long as Zen (although only half as long as Zen across the CCX boundary). There's some really good bits in Zen.

It doesn't explain why Windows 7 still seems to do a better job though with games.