Hardlimit test bank

[cobito]

Apart from the aesthetic changes and the possibility to sort search results based on both single and multi-threaded results that were introduced yesterday, today it is already possible to consult information about individual processor models. To do this, you just have to enter the link of the model that will appear in the tops and in the search results.

The link will only be available when there is additional information about the model. On the page of each model, in addition to the additional information, all the results sent for that model are shown.

As more results arrive, in the future on the pages of each model, performance statistics, graphs of distribution of results, etc., will appear.

By the way, submissions for different modes are also welcome. If your micro can SSE3, having results of all lower modes (FPU, MMX, etc., but especially of Mode 0 [FPU]) is also useful. Not everything in life is AVX2 and in order to be able to make comparisons with real-world results, it is good to have this type of information.

[kynes]

@cobito Check out the latest result I uploaded, you can cheat on the bench.

[cobito]

@kynes Apart from posting 32 threads, have you done anything else?

[kynes]

@cobito No, únicamente eso.

[cobito]

@kynes Well that was expected. I will investigate it.

By the way, I will end up deleting that result.

[kynes]

@cobito Damn, I can't use it as a signature anymore

[cobito]

The problem seems to be due to a failure in the synchronization of the execution of the processes. It appears mainly on slower PCs. For the moment I have patched the issue by limiting the maximum number of threads to at most twice the detected threads. In order to continue sharing results, you need to download the new version.

[Xevipiu]

My best friends are inside the Hwbot Staff, I can pass them the application so they can take a look

[cobito]

@Xevipiu That would be great. The more of us that participate, the more useful it will be for everyone when it comes to making comparisons.

[Xevipiu]

What I miss is a process LOG interface, so the wait wouldn't be so boring

[cobito]

@Xevipiu Tomo nota.

I have rewritten the frequency meter from scratch. If someone with OC would be so kind as to test it, I would appreciate it.

[Fassou]

@cobito Better, but it shows the serial clock, rather than the active

[cobito]

There is a new version of the program that, this time yes, should detect the actual frequency of the processor.

[Fassou]

@cobito It's clear that my processor is a rebel

That said, it takes about ten seconds to close the program, thinking :thinking:

[Sylver]

@Fassou Well blessed are you. It took me about a minute or more

[cobito]

The frequency meter has a small error that will be corrected. Regarding the boot time, yes, it needs to be improved. For now, it should take around 10 seconds.

[Xevipiu]

In processors that are not 100% natics to AVX2, their execution would have to be capped, since it sounds like a hedge the score

[cobito]

@Xevipiu What do you mean by processors that are not 100% native to AVX2?

[Xevipiu]

@cobito said in Hardlimit test bench:

@Xevipiu What do you mean by processors that are not 100% native to AVX2?

There are processors that do not use the instructions 100%, but when they do, they give result inaccuracies.
But leaving this aside, using AVX vs AVX2, the results are abysmal.

You would have to cap them according to the instructions of each processor to go well, don't you think?

[cobito]

@Xevipiu I still don't understand. What do you mean by there are processors that don't use instructions at 100%? The benchmark is designed to do 4 tests: two of the SIMD type (where having AVX for example will improve the results), one of memory (where more cache and/or faster memory will give you better results) and a generic test whose result has little dependence on the instruction set.

If you want to compare two processors under equal conditions, you just have to pass the benchmark in the same mode for both. If you don't have a result in the same mode, you can look at test#4 to get an idea.

The operations that are executed in all modes are exactly the same. If in a processor in FPU mode it gives you 1000 points in test#1 and changing to AVX2 mode it gives you 15000 points it simply means that for the vector calculation of integers, AVX2 multiplies by 15 in performance. That is the reality. Another matter is that then you are not going to run multimedia software or games and you don't care about vector instructions. In that case, once again, you either have test#4 or the complete benchmark in a lower mode.

One of the objectives of HLBM is precisely to evaluate how much the recent sets improve performance and also how optimized the older sets are in current CPUs.

As I understand it, what you are asking is that I weigh the results of higher modes so that the final result is similar to the execution in a lower mode? In that case it wouldn't make sense to have the possibility of evaluating the performance of recent sets.