Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
pretty good Kaap. Did you try the memory at 1100mhz? And 1700mhz on the core still pretty decent, I've seen dozens of AIO users on other forums not even hit 1700mhz and remain stable. I stick by what I've been saying, a Vega64 that can do 1700mhz and up, and remain at those clocks through benching or gaming, appears to be a pretty good performer. I still think as these drivers come along, it will pull well clear of the 1080.
It will only do 1100mhz when the card is still cold just after boot up.
Most of them are already pushing their limits. AMDMatt has some sort of anomaly card.It is a very poor overclocker on the core but the memory can do 1080.
Core voltage and clock impacts the HBM so as you push higher clocks or more voltage it could force your HBM overclock to become unstable. IMO, get the HBM as high as you can then work on the core. It's possible that lowering the core clock or undervolting the core will help you achieve a higher stable HBM clock and improve performance. You'll need to find the balance then test.Sapphire AIO owner here.
Without adjusting any voltages I can get 1050MHz on my HBM and 1770MHz core whilst running a +50% power increase; this gives a timespy score of 7932, anything higher results in a black screen crash.
I'm a bit confused about what to do with the voltage. Do I undervolt to reduce temperatures which results in improved stability, or overvolt to increase stability but will obviously increase temperatures?
Here is my AIO running Firestrike Standard.
Score 21 650
With AMD Radeon RX Vega 64(1x) @1700/1090
And Intel Core i7-6950X @4.444
Graphics Score 25 185
Physics Score 27 592
Combined Score 9 114
17.9.1 Drivers
https://www.3dmark.com/fs/13606407
It does ok on this bench because of the CPU clockspeed.
Core voltage and clock impacts the HBM so as you push higher clocks or more voltage it could force your HBM overclock to become unstable. IMO, get the HBM as high as you can then work on the core. It's possible that lowering the core clock or undervolting the core will help you achieve a higher stable HBM clock and improve performance. You'll need to find the balance then test.
Sept - Oct ? I figured it was going to take this whole year for Vega to level out, nano this year I dont knowWhere are the 3rd party cards?
Sept - Oct ? I figured it was going to take this whole year for Vega to level out, nano this year I dont know
Sapphire AIO owner here.
Without adjusting any voltages I can get 1050MHz on my HBM and 1770MHz core whilst running a +50% power increase; this gives a timespy score of 7932, anything higher results in a black screen crash.
I'm a bit confused about what to do with the voltage. Do I undervolt to reduce temperatures which results in improved stability, or overvolt to increase stability but will obviously increase temperatures?
Good advice this from Kundi. Two of my 64's top out at 1075Mhz, one does all the way up to 1100 but it artifacts. The other crashes past 1075Mhz. My best sample does 1100Mhz at stock voltage.Core voltage and clock impacts the HBM so as you push higher clocks or more voltage it could force your HBM overclock to become unstable. IMO, get the HBM as high as you can then work on the core. It's possible that lowering the core clock or undervolting the core will help you achieve a higher stable HBM clock and improve performance. You'll need to find the balance then test.
No RMA ?
+1It does what it says on the box so no reason to RMA the card.
It does no one any favours when people RMA cards because they have not won the silicon lottery in the overclocking stakes.
+2.It does what it says on the box so no reason to RMA the card.
It does no one any favours when people RMA cards because they have not won the silicon lottery in the overclocking stakes.
PowerColor AIOs are in, £780
Temperature isn't is a big a factor with the AOI cards and can easily be kept away from the 70 degrees throttling limit by increasing the fan RPM. Unless you have the card under a custom water loop and are achieving temps below 50 I don't think there are any significant performance gains between the 55 and 65 your card is likely to mostly sit between at full load. As AMDMAtt said, pushing the voltage in state 7 a bit will allow you card to hit the highest boost clocks although you will likely see the clock frequency fluctuate quite a bit, for me this is usually between 1670 and 1730. Lowering the voltage and clock frequency allows for a more stable boost clock, I was able to achieve a more solid boost clock 'line' in the Wattman monitoring graph, staying much closer to 1700 although the actual performance between the higher voltage and clocks versus lower voltage were largely on a par. Actually I found that the previous drivers favoured the lower voltages more than 17.9.1. Here are the details of results I've achieved at various settings and drivers.Sapphire AIO owner here.
Without adjusting any voltages I can get 1050MHz on my HBM and 1770MHz core whilst running a +50% power increase; this gives a timespy score of 7932, anything higher results in a black screen crash.
I'm a bit confused about what to do with the voltage. Do I undervolt to reduce temperatures which results in improved stability, or overvolt to increase stability but will obviously increase temperatures?
Downloaded Battlefield 1 last night. Best performance for me was 1702/1150 and 1777/1200 on states 6 & 7 (power limit and memory are always +50% and 1080 for me now) - boost clock stayed fairly close to 1710.interesting kundi.
Here's that video anyway, got a good solid 1750mhz overclock now