Announcement

Collapse
No announcement yet.

More AMD A10-7850K Kaveri APU Linux Benchmarks

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • edgar_wibeau
    replied
    No, it works, but e.g. A10-7850K clocks down to 2.4 GHz when set to 45W which is way below A8-7600's 3.1 GHz.

    That guy is an editor for Golem.de, a german tech site and he was involved in their Kaveri review.

    Leave a comment:


  • AJSB
    replied
    Originally posted by edgar_wibeau View Post
    I've read somewhere that it's possible to set the TDP for the higher rated APUs too, but AMD only guarantees that it works as expected for the A8-7600 for the time being.

    You can see here that in this UEFI Setup there is a mulitplier of 41 set (resulting in 4.1 GHz), so it mast be a "K" Kaveri - either a 7850K or a 7700K. And at least it's possible to set a lower TDP (65W or 45W) value or leave it at default.

    Yep, I'm also looking forward to the 65/45W cTDP - and even more so for the mobile - parts!
    Yeah that would be cool if it was that way....but i didn't see it in any review of the 2 K versions announced...it would mean:

    Disabled (AKA 95W for A10-7850K and A10-7700K)
    45W
    65W

    Leave a comment:


  • lovenemesis
    replied
    Future investment

    As an owner of A10-5800K with R9 270X, there's no immediate reason for me to jump on A10-7860K now. As it shown in these benchmarks, the improvement of Kaveri over Trinity on CPU side isn't that huge. No surprise.

    However, people should be notice that none of these benchmarks touches the true feature highlights of Kaveri: HSA/hUMA.
    The time when HSAIL backend shown in LLVM stable release is where power of Kaveri unleashed.

    Also, TrueAudio and Mantle are another two highlights for Kaveri, but I doubt they can make their way to Linux world.

    Leave a comment:


  • guido12
    replied
    For me, as long as there aren't crashes and weird display artifacts (ie. occasional flicker, weird blocks and other visual issues) then the performance/dollar will determine if I go with a <= 65 W Kaveri or Haswell chip in my upcoming mini-ITX PC. Can't wait for the CPU comparison benchmarks with Haswell.

    Leave a comment:


  • juanrga
    replied
    Second try.

    OpenBenchmarking.org, Phoronix Test Suite, Linux benchmarking, automated benchmarking, benchmarking results, benchmarking repository, open source benchmarking, benchmarking test profiles


    OpenBenchmarking.org, Phoronix Test Suite, Linux benchmarking, automated benchmarking, benchmarking results, benchmarking repository, open source benchmarking, benchmarking test profiles






    x264. I predicted 102 for Kaveri @4GHz and Kaveri @4GHz scores 94.45. Percent error = -7%

    Kaveri at stock scores 90.34. Percent error = -11%

    JTR. I predicted 4310.46 for Kaveri @4GHz and Kaveri at stock scores 3631. Percent error = -16%

    Kaveri @4GHz would score ~3925. Percent error = -9%

    C-Ray. I predicted 38.96 for Kaveri @4GHz and Kaveri @4GHz scores 37.00. Percent error = -5%

    Kaveri at stock would score ~39.77. Percent error = ~2%

    Himeno. I predicted 845.64 for Kaveri @4GHz and Kaveri @4GHz scores 958.10. Percent error = +13%

    Kaveri at stock would score 886.24. Percent error = +5%

    If I did no mistake the average of the percent errors is -5% (for stock). Would I say that Kaveri CPU performs about 5% poor than I predicted?

    Leave a comment:


  • juanrga
    replied
    Originally posted by curaga View Post
    Isn't Kaveri bdver3, ie Steamroller?
    Yes, you are right.

    Leave a comment:


  • jasonditz
    replied
    Am I alone in thinking the Kaveri's benchmarks are a bit disappointing? I did a comparison to my own system

    OpenBenchmarking.org, Phoronix Test Suite, Linux benchmarking, automated benchmarking, benchmarking results, benchmarking repository, open source benchmarking, benchmarking test profiles


    And especially on compile times the new APU doesn't seem to offer any measurable benefit over my Richland.

    Leave a comment:


  • curaga
    replied
    Originally posted by juanrga
    Could it could be a compiler regression for bdver2 and the score improve in a posterior test with an updated compiler?
    Isn't Kaveri bdver3, ie Steamroller?

    Leave a comment:


  • AJSB
    replied
    Originally posted by Kemosabe View Post
    This was not possible with Richland, Trinity and Llano. The chances are low Kaveri will do it better.
    (Why should one want this by the way? O_o To use a spare unneeded K version as HTPC? Bad reason)
    If you read the review at AnandTech you can see that a A8-7600 with cTDP set to 45W beats the crap out of a stock A10-5800K in many gaming tests when....and the gaming performance difference to a A8-7600 at 65W or a A10-7850K at stock settings is many quite small and a perfectly acceptable....but in some games, the difference is much bigger...512 shaders of 7850K matters.

    Why i want to try to do something similar with a A10-7700K or a A10-7850K ?
    Because two reasons:

    1) I can scale performance UP if i desire to play a specific game with specific graphical settings that a A8-7600 simply won't cut it because it's locked in 1st place.

    2) I can scale it DOWN even further in some circunstances in a different way thanks to have unlocked multipliers....and with that reduce power drain.

    The problem of the A8-7600 is that if i want to OC it, i can't...the A10-7700K give us that choice....but can it be set in one way, with cTDP, or another, with undervolting and underclocking to achieve similar values ro a A8-7600 ?

    Actually a A10-7700K is a A8-7600 with unlocked multiplier (besides slight bigger CPU base frquency)....i don't see any other differences except if cTDP is disabled in the 7700K...

    The review at anandTech is not clear (in fact gives the impression of the opposite) about the A10-7700K and the A10-7850K can be set via cTDP....the only thing they done was undervolt and underclock slightly the 7850K and i liked the results no matter were clearly worse (power drain wise) than a A8-7600...but then again they didn't push further down than 3.5GHz the 7850K...i wonder if we could go with it down to 3.2GHz and disable Turbo Boost (this way is right in the middle of a A8-7600 with TB (3.1 <-> 3.3GHz) for example and what would be the power drain then.

    Notice i also like to undervolt iGPU...no bad effects from that and lot's less power drain and less heat to dissipate...

    Using a A10-7850K will give the benefit of 512 shaders vs 384 of the A8-7600...

    Leave a comment:


  • juanrga
    replied
    Excellent numbers! Thanks by benchmarking

    My original predictions about Kaveri were






    Comparing to obtained scores,

    x264: I predicted 102 and Kaveri scores 94.45. Percent error = -7%
    John The Ripper: I predicted 4310.46 and Kaveri scores 3631. Percent error = -16%
    C-Ray: I predicted 38.96 and Kaveri scores 37.00. Percent error = -5%
    Himeno: I predicted 845.64 and Kaveri scores 958.10. Percent error = +13%

    The average of the percent error is -4%. I am waiting for more scores.

    My predictions assumed a 4GHz CPU. Kaveri is finally 3.7GHz. If I had know the real frequency before, I would offer better estimations. For instance, I would predict 94.35 (=102x3.7/4) for x264. The percent error had been reduced to 1% then :lol:

    The John The Ripper score is too small. Part of the discrepancy can be explained by the clock reduction in Kaveri, but still there is about a 9% less performance. Could it could be a compiler regression for bdver2 and the score improve in a posterior test with an updated compiler?

    At the other hand, the Himeno score is much better than I expected. The improvement is even more impressive when we consider the clock reduction. The improvement is of 22%, which implies more than 30% IPC. This huge IPC gain is about the same found in Wprime, in games, and in Floating Point GeekBenches. Apparently the FPU units were not improved. I believe that the improvement in floating point performance is due to the improvements made in the L2 cache. Do you agree on this analysis?

    The average of the percent error is -4%. Would I say that Kaveri CPU performs about 4% poor than I predicted?

    Leave a comment:

Working...
X