AMD - Piledriver chitchat

[CAT-THE-FIFTH]

Bristol Ridge launching at Computex:

http://videocardz.com/59947/amd-to-l...10-at-computex

[DanceswithUnix]

WCCFtech video does a comparison run of the FX8350 vs the more expensive i5 6400 running the new DooM with a 980ti, and the FX wins on average and minimum frame rates.

https://www.youtube.com/watch?v=XZ6VAbHxzxM

[CAT-THE-FIFTH]

According to an AMD PR guy Zen will compare favourably to Skylake:

http://www.itwire.com/it-industry-ne...l-skylake.html

[CAT-THE-FIFTH]

Multiple Polaris entries in the GPUOpen database:

http://phoronix.com/scan.php?page=ne...laris-More-IDs

[DanceswithUnix]

Multiple Polaris entries in the GPUOpen database:

http://phoronix.com/scan.php?page=ne...laris-More-IDs

That's a lot of PCI IDs, guess you need to allow for the Polaris 11 variants with 16GB of super slow DDR3 ram

[shaithis]

WCCFtech video does a comparison run of the FX8350 vs the more expensive i5 6400 running the new DooM with a 980ti, and the FX wins on average and minimum frame rates.

I read somewhere recently that the newer AMD drivers now have *most* of the DLLs compiled with the non-Intel compiler, removing those pesky "compiler optimisations"

[watercooled]

I read somewhere recently that the newer AMD drivers now have *most* of the DLLs compiled with the non-Intel compiler, removing those pesky "compiler optimisations"

That's interesting - I would've thought AMD of all people would have been careful to watch out for things like that*? Maybe the management shake-up has prompted the change?

*Then again, The Stilt managed to significantly improve SuperPi (x87) performance with a microcode mod IIRC. http://forum.hwbot.org/showthread.php?t=78490

[CAT-THE-FIFTH]

Saw this in the OcUK Polaris thread:

There were some interesting bits of yesterdays AMD webinar:

Those two GPUs will have surely have their cut down versions and will have a wide range of different board designs"

"AMD also pointed out one of their key features in their Polaris architecture which is Asynchronous Compute."
"AMD is very keen on pushing this feature into upcoming titles and they stated they will drive it even more aggressively than before."
"they specified that they were able do get to get more performance out of each transistor not only thanks to 14nm FinFET process but also by virtue of the various optimizations within the Polaris architecture itself."

"We asked them if the efficiency improvements that AMD specified in their previous presentations are uniform for all GPUs AMD will offer or if it’s just for the most efficient GPU of the whole family. We were lucky enough to get my question answered with a little tiny bit of extra unexpected spice. AMD answered that there will be wide range of different board designs which means some will have different clocks than others. This means that some will be more efficient than others but generally the capability and improvements will be there."
"AMD mentioned AIB partners and that some of them will most likely try to drive the clocks as high as possible with their own designs, which would mean lower efficiency."
"AMD is not limiting AIB partners with clockspeeds for the sake of better efficiency"

This might mean they were mainly aiming for efficiency not performance, but AIB partnets might do the opposite. The former mention of 175w tdp but only 120-130w real power consumption pointed to the direction that they have room for improvement.

[DanceswithUnix]

I read somewhere recently that the newer AMD drivers now have *most* of the DLLs compiled with the non-Intel compiler, removing those pesky "compiler optimisations"

That would be interesting, though I wonder which compiler they have switched to. In the past people have found that even with the Intel shenanigans in the compiler, the Intel compiler has still been faster than the competition on AMD hardware. It is just that it could be faster still.

[CAT-THE-FIFTH]

Polaris 10 samples apparently running at 1.266GHZ:

http://www.guru3d.com/news_story/amd..._1_27_ghz.html

AMD conference on the 1st of June covers both Bristol Ridge and Polaris:

http://wccftech.com/amd-announces-ju...CCFtech.com%29

[CAT-THE-FIFTH]

Have people seen this:

https://www.reddit.com/r/pcmasterrac...oyuc?context=3

Should have gotten a 490x

RemindMe! 8 days

So is something happening on the 26th??

That date coincides with the Macau event from the 26th to the 29th of May:

http://wccftech.com/amd-polaris-10-p...-launch-event/

It appears the guy is an XFX person:

https://linustechtips.com/main/profi...00&type=status

[DanceswithUnix]

This isn't directly AMD related, but of interest to see what sort of underlying technology will be available to them for their upcoming CPUs.

Charlie has an interesting article on that ARM 10nm test chip recently with a lot of figures in it. The idea of a 3GHz successor to the A72 sounds very nice indeed.

Scary numbers on the cost of new designs though: "A design that cost about $5M on 28nm is said to rise to a bit over $30M on 10nm". I do wonder if when we will start seeing SOI 22nm chips coming out to get 16nm performance without the costs.

http://semiaccurate.com/2016/05/18/a...mis-test-chip/

[CAT-THE-FIFTH]

Saw this posted on Reddit:
http://phx.corporate-ir.net/External...92964179233292

Silicon running in labs now, meeting expectations and remain on track for availability in 2016

7th Gen APU ‘Bristol Ridge” delivers 20% graphics and CPU performance uplift

[jigger]

20% on Kaveri with the same overclocking potential would be a potent CPU.

[flearider]

do you think all this will be out by july .. ? it's my 50th and i'm going to do one last big build so new everything .. after that can't see me needing anything else for quiet some time

[scaryjim]

Saw this posted on Reddit:

Best footnote ever in that image: "see footnotes". Seriously? A footnote just to tell you to read the footnotes?! Come on AMD, that's just ridiculous...

EDIT: flearider - the only thing likely to be out for July is Bristol Ridge, which is an evolution of the bulldozer-style APUs. AFAIK AMD's new high performance core, Zen, isn't due until near the end of the year. If you're looking for a high performance build Bristol Ridge probably isn't for you. Indications are that Zen's performance will be generally competitive with Intel, rather than being workload specific (after all, bulldozer and its revisions are actually very good at certain workloads). So it's a question of if you're willing to wait, really.