Jump to content
HWBOT Community Forums

Recommended Posts

  • Leeghoofd pinned this topic
  • 3 months later...
Posted (edited)

I was wondering why the overall score including CPU is used for TimeSpy.

Wouldn't a gpu score only category be more suitable to compare video cards? Should be no problem if we still run the cpu test and not allow custom runs.

 

The current TimeSpy scores are a system score and do not really belong into only the "videocards" or "processors" category. A 3090 paired with a 6 core could score lower than a 2080 with 16 core cpu for example.

Edited by sonicSalad
  • Crew
Posted

And if we only count only the GPU score, you sure it does not scale with processor architecture? A 3D scores will only be pure GPU IF the tested GPU is fully maxed by the benchmark and/or we impose low CPU clockspeeds

Posted

Correct, this would not eliminate cpu bottlenecks and the same gpu will still perform different depending on the cpu model and tuning.

The intention is to have a clearer view of gpu performance within these two benchmarks. When going for top gpu score it is common to run 1ccx no smt and ignore cpu score.

This results in a couple extra gpu points and is the way some of the top gpu scores are achieved. This is also interesting for firestrike.

 

Would still be understandable to not add an extra category to prevent a split. Maybe adding seperate overall, cpu and gpu fields to the existing category would be more sensible. This way it would be similar to 3dmarks result page where you can choose combined/gpu/cpu.

 

OT.

probably has been discussed already, cinebench results with deactivated ccx (5900x with 6c/12t) will not be counted as 6 core cinebench result by rules or bench mate limitations? Example: https://hwbot.org/submission/4742496_sonicsalad_cinebench___r23_multi_core_with_benchmate_ryzen_9_5900x_11203_pts

Posted
42 minutes ago, sonicSalad said:

 

probably has been discussed already, cinebench results with deactivated ccx (5900x with 6c/12t) will not be counted as 6 core cinebench result by rules or bench mate limitations? Example: https://hwbot.org/submission/4742496_sonicsalad_cinebench___r23_multi_core_with_benchmate_ryzen_9_5900x_11203_pts

No. It is a 12c/24t processor.

Not allowed to core down to fit into another category. Already in the general rules.

  • Like 3
Posted (edited)

Thanks!

 

Is it allowed to load valid results from the autosave folder and take the screenshot some time later?

My best overall score was from before I used hwbot. The hardware is the same except a second radiator.

Smells like forging a result, but would be cool if it counts.

6800 1st

https://www.3dmark.com/spy/19329822

current ranking

https://hwbot.org/benchmark/3dmark_-_time_spy/rankings?hardwareTypeId=videocard_3325

Might be able to recreate it with older drivers etc.

Luckily this is the only result where I left info in the description, that should help. :cool:

 

edit:

just beat it! https://hwbot.org/submission/4745661_sonicsalad_3dmark___time_spy_radeon_rx_6800_17198_marks?recalculate=true

Edited by sonicSalad
  • 5 months later...
  • 7 months later...
Guest TonyBombassolo
Posted

Is there a process or a history that explains why HWBot has an inconsistent application of allowing LOD/Tess manipulation on benchmark submissions? I can understand legacy benchmarks like 3DMark03, but I am talking about current, supported and active benchmarks. 

Specifically, ALL of the Firestrike benchmarks have LOD/Tess explicitly allowed

For Time Spy, LOD/Tess manipulation is explicitly disallowed

Port Royal explicitly disallows LOD/Tess manipulation

To me it seems like Fire Strike submissions allowing LOD/Tess are inconsistent with that allowance for other "current-world" benchmarks (in this case from the same vendor and the same suite of products) and Im failing to understand why that is. Typically HWBot is pretty consistent with benchmark rules, so this seemed odd to me. One of my favorite things are the new updates for Benchmate, because in general it removes "software skill" from the "overclocking skill" point system by allowing everyone to play on the proverbial same field. Any thoughts on why we would have inconsistent LOD/Tess rules on GPU benches, especially ones that give Global points?

Posted
9 hours ago, TonyBombassolo said:

Is there a process or a history that explains why HWBot has an inconsistent application of allowing LOD/Tess manipulation on benchmark submissions? I can understand legacy benchmarks like 3DMark03, but I am talking about current, supported and active benchmarks. 

Specifically, ALL of the Firestrike benchmarks have LOD/Tess explicitly allowed

For Time Spy, LOD/Tess manipulation is explicitly disallowed

Port Royal explicitly disallows LOD/Tess manipulation

To me it seems like Fire Strike submissions allowing LOD/Tess are inconsistent with that allowance for other "current-world" benchmarks (in this case from the same vendor and the same suite of products) and Im failing to understand why that is. Typically HWBot is pretty consistent with benchmark rules, so this seemed odd to me. One of my favorite things are the new updates for Benchmate, because in general it removes "software skill" from the "overclocking skill" point system by allowing everyone to play on the proverbial same field. Any thoughts on why we would have inconsistent LOD/Tess rules on GPU benches, especially ones that give Global points?

It basically has to do with when systeminfo was able to actually detect it. Because there's lots of legacy scores where it could be used and not detected its still allowed as a tweak on those benches. By the release of Timespy and later system info had fixed this

  • Crew
Posted

Simple; new owner new rules ? 

Joke aside, from TimeSpy on, the previous moderation team adhered to Futuremark's standards, thus the run needs to be checked as valid on their site.  Initially not even a screenshot was required, but the systeminfo couldn't keep up with the creative software clockers so we reinstated that standard screenshot rule for TimeSpy...

If we would force the popular Firestrike series to no longer use LOD we would get even more upset overclockers as they can't beat older records... we imposed VALID runs for cloudgate and co which were less popular...  you can't imagine the hate mails I got...

 

  • Like 1
  • Thanks 2
Guest TonyBombassolo
Posted

Thanks everyone, I knew there was some logic behind it :)

  • 1 year later...
Posted (edited)

Image, you would want to post one of your 3DMark results here, but those results were validated before you registered here. On the 3DMark site, you uploaded your validation if you chose to, and they validate, when everything is according to their liking.

So even though the results very officially validated by them, you cannot post them here, because obviously the 3dMark client was closed during the time of the validated result and the registration here, because you cannot produce the screenshot with your client, showing the "valid" or not.

And to make matters even worse, 3DMark does offer you to connect your 3dMark account and the HWBOT account, but HWBOT doesn't import your 3DMark results, nor does HWBOT register, that the two accounts have been linked in the 3DMark client.

So posting my 3DMark results, even though they are all valid, is a no no here?

Edited by elita
  • Crew
Posted

FYI 3Dmark benchmark program does Auto saves, so you can load the result in the program, open  2 x CPUZ tabs (CPU & Memory) and  one GPUZ tab and make a screenshot. For air and water cooled setups it's easy to (re)do, for extreme cooled setups its more time consuming.

Why do we want screenshots:

Submitting with screenshots provides way more data than a verification link, it is also easier to verify and compare with a screenshot and to make matters even worse not all VALID scores are really VALID. 

A quick summary of competitive benchmarking: to get a better score and ranking users tend to bend rules, the build in software detection is not always top notch and can be tricked. Therefore some of the top scores Validated  by 3DMark at UL still require once in a while manual moderation as this community is very creative. 


Anyways we were contacted by Ari from UL that there is an issue due to the new code thaét was implemented a few weeks back. Our programmer is waiting on more data from UL to fix it.


Only thing you need to do is to make a proper verification screenshot after each score  according HWBOT rules , it takes 5-15 seconds. Without this mandatory screenshot your scores might get reported.

Posted

Sorry, ich kann leider keinen Verifizierungslink liefern, leider habe ich keine Treiber für meinen Z690 unter Win 7, obwohl ich einfach mal alles im default laufen lasse, oder einwenig OC, wie immer Matti hat immer die Arschkarte 

 

Sorry, unfortunately I can't provide a verification link, unfortunately I don't have any drivers for my Z690 under Win 7, although I just let everything run by default, or a little OC, as always Matti always has the ass card

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...