Photo Benchmark 11021 Results

[McD]

& for Walt on the PC forum (AMD R9-5900 + RTX3090)...

 

[MikeTO]

I'll update the table one last time after somebody posts data from a new M2 MacBook Air or MacBook Pro.

[McD]

On 7/6/2022 at 1:58 AM, MikeTO said:

I'll update the table one last time after somebody posts data from a new M2 MacBook Air or MacBook Pro.

And from another MaxTech review for the 10-core M2 MBP. Nice single core CPU but not quite the 12900K - I wonder if it uses 2xNEON or just 128-bit.  Odd that the combined score is higher than the Raster score, I don't think I've seen that before.

Edited July 6, 2022 by McD
Added "combined score" observation

[debraspicher]

[MikeTO]

One last update

[debraspicher]

On 9/11/2022 at 9:26 AM, MikeTO said:

One last update

@FrankVoss 5900X kicking ass... also a 3060 TI nearly beating a 3090... lol

[McD]

On 9/16/2022 at 3:13 AM, debraspicher said:

@FrankVoss 5900X kicking ass... also a 3060 TI nearly beating a 3090... lol

Yeah & the ‘combined’ scores are up there with the MacBook Air.

[Pandovski]

I'd rather get the 3060 over the 3060 Ti or the 3070 because of the extra 4GB of VRAM. I've seen 8GB of VRAM slurped just by Affinity at times. When I used the GT1030, the problem was never the lack of graphic processing power, it was always the 2GB VRAM bottleneck. 

This is a single fan 3060 with a 5.1GHz OC 10700K. 

[debraspicher]

6 hours ago, Pandovski said:

I'd rather get the 3060 over the 3060 Ti or the 3070 because of the extra 4GB of VRAM. I've seen 8GB of VRAM slurped just by Affinity at times. When I used the GT1030, the problem was never the lack of graphic processing power, it was always the 2GB VRAM bottleneck. 

This is a single fan 3060 with a 5.1GHz OC 10700K. 

 

I actually went 12GB precisely because of what you mention. It seems more VRAM is better when it comes to Affinity programs. I'm sure in the future as the GPU is utilized more by design/illustration apps, it would then make more sense to have more than enough VRAM on hand for multi-tasking, especially running different software side-by-side... which is often for some.

The GPU scores seem strongly correlated to CPU, at least on a quick scan. This may have been discussed prior on the forum but I would've missed it. I know it was mentioned earlier in the thread that GPU<->CPU memory speed/bandwidth plays a role which gave Apple M1's a leg up. So...

I've been planning an update for my processor to finish out AM4 probably when AM5 drops. So I guess I'll find out on my machine anyway when I can rerun the benchmark.

[Klaus Lotz]

Hi all, here is Benchmark from Ipad Pro, 3.Gen., A2459 with M1:

 

[debraspicher]

On 8/9/2022 at 7:03 PM, debraspicher said:

 

New results with CPU change (3600 to Ryzen 9 3900X) .. RAM speed bump next. GPU score also up.

[debraspicher]

Husband took my old CPU and ran benchmark

(Edit) Windows 10 (Build 19043)
Mobo: ASUS Rog Strix B-450-F Gaming
CPU: Ryzen 5 3600 @ 3.6Ghz
RAM: 16 GB 3200 (DOCP Enabled)
GPU: Gigabyte Nvidia 3070 8GB OC Edition
Monitor 1: 1440p @ 110%

[debraspicher]

CPU change(s) again! (x2)

Microsoft Windows 10 Home (Build 19043)
AMD Ryzen 7 5800X @ 3.8Ghz; Mobo: Asus X470 Prime Pro
32GB DDR4 (3000Mhz)
EVGA Nvidia GeForce GTX 3080 X3C Ultra 12GB
Monitor 1: 4K @ 150%
Monitor 2: 4K @ 150%

Benchmark:

[debraspicher]

CPU change in the other machine as well:

Windows 10 (Build 19043)
Mobo: ASUS Rog Strix B-450-F Gaming
CPU: Ryzen 5 5600x @ 3.7Ghz
RAM: 16 GB 3200 (DOCP Enabled)
GPU: Gigabyte Nvidia 3070 8GB OC Edition
Monitor 1: 1440p @ 110%

[MikeTO]

Okay, once last update to add in all of Debra's upgrades and Klaus' iPad.

[chrisb123]

5950x

and 6900xt that does nothing in affinity

[debraspicher]

I have not done extensive tweaks, just a simple All Core > -20 undervolt to the 5800x. Stable 4.5-4.6ghz in cinebench r32 in multi-core... same boost 4.85ghz. I did not uncap it in the BIOS.

I wouldn't add to chart. It's just to show benefits of undervolting in Ryzen. Very competitive for my workload against the 12c chips... also showing once again just how much base clocks help with GPU-accelerated functions/scoring...
 

.

[debraspicher]

Pretty good results. The right is the "best" one so far. 22K in GPU...

I undervolted to -30 (max) since last run and with testing, seems stable. Bumped boost clock +200mhz to 5ghz.

[debraspicher]

There is no beta version yet, but if anyone is curious about the V2 benchmark, it has been updated to 20000 (lol).

This result is with my overclocks, so it doesn't compare to @Miketo's chart results, but would compare the most to the third column in my most recent post. Although I've changed my RAM to a higher speed recently. That did not really seem impact results really, so I did not post any updates.

I noticed right away the single GPU results are much lower. Did they update the test to rely less on single core performance when testing GPU? We won't know until more people test if it still scales with single core or not.

[walt.farrell]

1 hour ago, debraspicher said:

but would compare the most to the third column in my most recent post.

Benchmarks between V1 and V2 are not comparable, according to what it says on the benchmark text. 

Second paragraph of text in the right in your screenshot.

[debraspicher]

10 minutes ago, walt.farrell said:

Benchmarks between V1 and V2 are not comparable, according to what it says on the benchmark text. 

Second paragraph of text in the right in your screenshot.

Yes, I know. I'm making note of the differences.

[Pandovski]

Here's mine, didn't change anything from previous setup (5.1GHz all-core), just installed the v2 Affinity suite. 

[ErikA310]

Here is also a replay of my measurements

Hardware changes:

• None,

Software changes:

• AP 2.0.0
• MacOS Ventura 13.0.1

(Converstion to text format is attached as well AP 2.0.0 Benchmark.txt)

 

[Duskstalker]

5900X and RX5700XT

[MikeTO]

I think it would be best to start a new thread titled "Photo Benchmark 20000 Results" and share these results there since they aren't comparable to 11021 results.