Announcement

Collapse
No announcement yet.

NVIDIA GeForce GTX 1070 Ti, Takes On The Radeon RX Vega 64 Under Linux

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

  • #11
    Originally posted by chuckula View Post
    2. The fact that the GTX-1070Ti is really only losing to the Vega 64 in OpenGL in Mad Max while easily beating the Vega 64 in Vulkan appears to put to rest the propaganda I've been hearing for years about how only AMD is capable of designing hardware that can use Vulkan.
    Are you seriously using Linux gaming performance to show how NVIDIA beats AMD on Vulkan?
    Man, you fanbois are desperate these days.

    (explanation: AMD is not really caring about Vulkan on Linux at this moment as they have higher priorities, it's not a secret that their windows driver is much better)

    Comment


    • #12
      Originally posted by starshipeleven View Post
      (explanation: AMD is not really caring about Vulkan on Linux at this moment as they have higher priorities, it's not a secret that their windows driver is much better)
      I'm sorry... didn't AMD invent Vulkan single handedly or something like that? Once again, I've been hearing empty marketing propaganda for YEARS that NVidia is full of idiots who literally can't design hardware that can use Vulkan at all. Here we see proof that this is of course nonsense. And your excuse is that AMD doesn't care about it's supposed invention of Vulkan?

      Well then OK: starshipeleven just said on the record that if you care about Vulkan on Linux you spend money on Nvidia hardware because according to AMD Vulkan isn't important.

      Comment


      • #13
        Wrong decisions by AMD. Me and a friend we flashed a Vega56 at ~1050mv, ~1.6Ghz and ~150wp. Well nothing comes near it in any benchmark except a 1080ti that also strangles vs our card.

        Michael you should bench something similar.

        Comment


        • #14
          Originally posted by chuckula View Post
          I'm sorry... didn't AMD invent Vulkan single handedly or something like that? Once again, I've been hearing empty marketing propaganda for YEARS that NVidia is full of idiots who literally can't design hardware that can use Vulkan at all. Here we see proof that this is of course nonsense. And your excuse is that AMD doesn't care about it's supposed invention of Vulkan?

          Well then OK: starshipeleven just said on the record that if you care about Vulkan on Linux you spend money on Nvidia hardware because according to AMD Vulkan isn't important.
          When you're done purposefully misreading my statements for the sake of trolling here, you can head to your favourite Windows benchmarks site and see that you're posting bullshit.

          Comment


          • #15
            I don't think customers are too interested in how well Vega will perform in 2 years time. Wish AMD would understand this. A 12.5TFLOP card being beat by a 8.1TFLOP card (1070ti) is never a pretty sight.

            Originally posted by artivision View Post
            Wrong decisions by AMD. Me and a friend we flashed a Vega56 at ~1050mv, ~1.6Ghz and ~150wp. Well nothing comes near it in any benchmark except a 1080ti that also strangles vs our card.
            While downvolting a VEGA card does make its average and low performance more stable and score better overall, I very much doubt it comes anywhere near 1080ti performance, perhaps you mean 1080. (or show us your benchmark results perhaps?). For a Vega to come close to a 1080ti it would need 2ghz core or 1.2ghz memory clocking. (HBM2 seems to top out at 1050mhz from what I hear)
            Last edited by theriddick; 07 November 2017, 09:11 PM.

            Comment


            • #16
              Originally posted by theriddick View Post
              I don't think customers are too interested in how well Vega will perform in 2 years time. Wish AMD would understand this. A 12.5TFLOP card being beat by a 8.1TFLOP card (1070ti) is never a pretty sight.



              While downvolting a VEGA card does make its average and low performance more stable and score better overall, I very much doubt it comes anywhere near 1080ti performance, perhaps you mean 1080. (or show us your benchmark results perhaps?). For a Vega to come close to a 1080ti it would need 2ghz core or 1.2ghz memory clocking. (HBM2 seems to top out at 1050mhz from what I hear)
              Random benchmarks: https://www.hardwareluxx.de/index.ph...ting-test.html

              Many Vega64 and RX480/580 have dirty silicon, prefer Vega56 and RX470/570, the laser cutter does better job.

              Comment


              • #17
                Originally posted by artivision View Post
                ....
                As expected, your wrong. Here's why, but I'm sure your contest all this, I won't argue or reply, this is just what I see as fact.
                (1) Its windows not linux,
                (2) Benchmarks are vs a Founders Edition 1080, try normal OC editions which are cheaper
                (3) Its still far from 1080ti speeds,
                (4) Even in these few results it still shows the Vega64 being better by a tiny bit, so I have no idea what your talking about silicon lottery here.
                (5) Vega56 is TDP limited and Vega64 cards can reach 550mhz (1050-1100) on HBM2 which is going to help allot better, plus better TDP allowance in firmware.
                (6) Yes price wise Vega56 is better, but performance and OC potential wise, no.

                Comment


                • #18
                  Originally posted by theriddick View Post
                  I don't think customers are too interested in how well Vega will perform in 2 years time. Wish AMD would understand this.
                  I'm pretty sure AMD does understand that. They were just under intense pressure to push out Vega even if it wasn't ready yet - and it wasn't. The alternative was to announce a delay and customers wouldn't have liked that either. At this point I'm just hoping they've applied the lessons learned for their next hardware generation. Raven supposedly has several improvements to it over Vega 1 so it's not just a copy of the architecture.

                  Comment


                  • #19
                    Originally posted by chuckula View Post
                    what a gulf there is in software support in Nvidia's favor
                    Or how bad amd drivers and architectures are for anything but mining in practice.
                    Last edited by xpue; 08 November 2017, 01:51 PM.

                    Comment


                    • #20
                      Originally posted by Michael View Post

                      I'll have OpenCL tests in a follow-up article.
                      Hi, I miss the FAHbench testing of the GTX 1070 Ti.

                      If you have a link to the benchmark results, I would be most thankful.

                      Best regards,
                      Ken

                      Comment

                      Working...
                      X