Post your GPU scores - to prep for next gen cpu's I have 1080ti 3080ti 6900xt included

[Skeeter]

https://community.amd.com/t5/graphics/r ... 604#M78542

ur post is visible now, maybe was just marked down as check before show kinda thing like on steam forums?

tho u seem to have forgot my fps results with my 6700xt.

[Panos]

https://community.amd.com/t5/graphics/r ... 604#M78542

ur post is visible now, maybe was just marked down as check before show kinda thing like on steam forums?

tho u seem to have forgot my fps results with my 6700xt.

Where you posted your results? I triple check the 3 pages cannot find them

Np found it.

The 6700XT should have been faster than 80 if I get 70 at spire.

[Eyeklops]

Btw if anyone reading this and has access to a Radeon VII which is comparable to the 5700XT please run the test to see if it is the vram bus or not.

My gpu is a Radeon VII. I will run the test later tonight. The architecture is GCN (same one used in Vega graphics cards). It differs substantially from RDNA used in the 5700 XT. They are somewhat comparable in performance though.

[Panos]

Btw if anyone reading this and has access to a Radeon VII which is comparable to the 5700XT please run the test to see if it is the vram bus or not.

My gpu is a Radeon VII. I will run the test later tonight. The architecture is GCN (same one used in Vega graphics cards). It differs substantially from RDNA used in the 5700 XT. They are somewhat comparable in performance though.

Yep. However has 10 times the VRAM Bus width

[Skeeter]

I have wondered, my 6700xt in gpuz says vram bandwidth is like 380 or something which isnt great vs say nvidia cards but i dont think that takes into account the 96mb infinity cache which i think effects the bw to be alot more when in use.

https://www.techarp.com/computer/amd-in ... explained/

I wonder if its used properly in x4?

[Panos]

I have wondered, my 6700xt in gpuz says vram bandwidth is like 380 or something which isnt great vs say nvidia cards but i dont think that takes into account the 96mb infinity cache which i think effects the bw to be alot more when in use.

https://www.techarp.com/computer/amd-in ... explained/

I wonder if its used properly in x4?

RDNA1 doesn't have Infinity Cache, yet the GTX1080Ti is 60% faster than the overclocked 5700XT. Something that is not the case in any other game.

Atm thinking to go and open the box to plug the Vega 64 as the PCIe3 8x to test it too. Cannot put it on the normal slot as I have watercooling.

[Skeeter]

I know, just wondered if it was a factor for the 6 series of cards or not.

[Eyeklops]

Btw if anyone reading this and has access to a Radeon VII which is comparable to the 5700XT please run the test to see if it is the vram bus or not.

Results from the Radeon VII
System Specs:
i9-9900KS @ 5.0GHz all cores
32GB 3600MHz RAM
NVMe Samsung SSD
CPU & GPU on water cooling and nowhere near thermal limits.

Lowest framerate was always the Honshu flyby.
2560x1440 144Hz Dual Monitor. Primary monitor for gaming, secondary for system monitoring.
72-108 FPS, ~195 watts GPU usage, ~1750MHz Core, 1200MHz vRAM - Increased vRAM speed and undervolted GPU.
75-113, ~280 watts GPU usage, ~1900MHz Core, 1200MHz vRAM - Heavy overclock with +20% power limit.
68-103, ~190 watts GPU usage, ~1750MHz Core, 1000MHz vRAM - Mostly stock settings except for an undervolted GPU.

I also tested with the CPU @ 5.2 GHz but there was no discernable difference in performance (as expected).

[Panos]

Btw if anyone reading this and has access to a Radeon VII which is comparable to the 5700XT please run the test to see if it is the vram bus or not.

Results from the Radeon VII
System Specs:
i9-9900KS @ 5.0GHz all cores
32GB 3600MHz RAM
NVMe Samsung SSD
CPU & GPU on water cooling and nowhere near thermal limits.

Lowest framerate was always the Honshu flyby.
2560x1440 144Hz Dual Monitor. Primary monitor for gaming, secondary for system monitoring.
72-108 FPS, ~195 watts GPU usage, ~1750MHz Core, 1200MHz vRAM - Increased vRAM speed and undervolted GPU.
75-113, ~280 watts GPU usage, ~1900MHz Core, 1200MHz vRAM - Heavy overclock with +20% power limit.
68-103, ~190 watts GPU usage, ~1750MHz Core, 1000MHz vRAM - Mostly stock settings except for an undervolted GPU.

I also tested with the CPU @ 5.2 GHz but there was no discernable difference in performance (as expected).

Thank you.
Yep, there is issue with perf on the AMD cards in general, regardless GCN or RDNA1/2.
Hoping Egosoft has some good data to look at the issue.

[Skeeter]

Dont forget to add the result to the amd thread at amd support forum.

I tried passing this and other issues to a amd rep at ocuk forums, never know might help.

https://forums.overclockers.co.uk/posts/35049727

[Panos]

Dont forget to add the result to the amd thread at amd support forum.

I tried passing this and other issues to a amd rep at ocuk forums, never know might help.

https://forums.overclockers.co.uk/posts/35049727

Matt will respond too. If not I will send him an email...... or 1000

Hope this is resolved otherwise need to start looking for a RTX3080 closest to MSRP as possible

(mirror started screaming SHAME, SHAME )

[Imperial Good]

Check AMD GPU performance running the latest beta 4.10 in case there were changes made.

I recall X4 having issues with AMD GPUs in the past, such as graphic bugs and the like. I think the devs mentioned working with AMD to try and solve them. I would not be surprised if the performance issues with the RDNA2 cards are caused by some sort of hack to fix those graphic bugs. This could range from vendor specific shader code to AMD turning off various driver level optimisations for X4.

[Panos]

Check AMD GPU performance running the latest beta 4.10 in case there were changes made.

I recall X4 having issues with AMD GPUs in the past, such as graphic bugs and the like. I think the devs mentioned working with AMD to try and solve them. I would not be surprised if the performance issues with the RDNA2 cards are caused by some sort of hack to fix those graphic bugs. This could range from vendor specific shader code to AMD turning off various driver level optimisations for X4.

My tests are with 4.10 beta 6 and both 21.6 and 21.8 drivers just to be sure and using RDNA1 not 2, so is the other user above using CGN-Vega (Radeon VII).
If needed I will try to see how to fit the Vega64 backup card I have, but need to work around the watercooling loop.

[Skeeter]

Always been on the latest betas of 4.10 for awhile, hence why ive been reporting seeing artifacts in them with 21.6.1 and 21.7.x and 21.8.1 drivers. I know x4 has had issues with glow causing artifacts in the past, yet when i report em they seem to not be bothered much.

But along with artifacts reported in other topics, we have this topic questioning amd performance in general for x4.

Something rather serious is going on with x4 and amd cards, either its amd drivers or ego not able to get amd cards working well with vulklan. Or a mix of the two.

[Eyeklops]

Same. Tests run with 4.10 beta 6 and either newest, or next newest AMD drivers.

[Skeeter]

https://forums.overclockers.co.uk/posts/35051996

An update of sorts from amd rep.

[Panos]

https://forums.overclockers.co.uk/posts/35051996

An update of sorts from amd rep.

Good one Skeeter

That I do not reply doesn't mean now following both the discussion

[Panos]

https://forums.overclockers.co.uk/posts/35051996

An update of sorts from amd rep.

Tested last night on windows with 21.8.2 just in case, no change.
I will test tonight on the fresh Manjaro installation, with latest MESA and 5.14.rc7 kernel

[Skeeter]

I tested 21.8.2 too.

Same.

I have tried chill for a change tho, i was sick of the wide disparity between 70fps in space (sometimes) and then on station (can sometimes find a lovely stable high fps station but often its fluctuate city, the varied stations, lol. 30fps - 50 fps (30 alot of the times, depends on the situation at the time, can go up and down like a yo-yo even when nothing seems to be happening). So i thought right the sudden switching of fps is driving me nuts, id rather a consistence frame time so i limited chill to 30-30, it wasnt great but 30-40 was alot smoother, not great but smoother experience than one min 30 then 35 then 40 then 29 then 45 then 35 then 47 etc.

Its a shame i do min install so i dont have game profiles which would be handy atm if i do use chill now for x4, if i did full install then id get em but then i get all the other crap like overlay, relive, and the likes yuk no thanks.

[Panos]

I tested 21.8.2 too.

Same.

I have tried chill for a change tho, i was sick of the wide disparity between 70fps in space (sometimes) and then on station (can sometimes find a lovely stable high fps station but often its fluctuate city, the varied stations, lol. 30fps - 50 fps (30 alot of the times, depends on the situation at the time, can go up and down like a yo-yo even when nothing seems to be happening). So i thought right the sudden switching of fps is driving me nuts, id rather a consistence frame time so i limited chill to 30-30, it wasnt great but 30-40 was alot smoother, not great but smoother experience than one min 30 then 35 then 40 then 29 then 45 then 35 then 47 etc.

Its a shame i do min install so i dont have game profiles which would be handy atm if i do use chill now for x4, if i did full install then id get em but then i get all the other crap like overlay, relive, and the likes yuk no thanks.

If I could find an AIB 3080 at MSRP (around £700) would have bought it because I doubt we going to see any AMD perf improvements for months.