Basemark GPU Wants To Benchmark Everything That Has A GPU

[Guest]

Basemark recently released its Basemark GPU benchmark with support for Windows, Linux, and Android.

Basemark GPU Wants To Benchmark Everything That Has A GPU : Read more

 

[urbancamper]

2.2 gb is a lot for a benchmarking tool I have never heard of. Quite frankly. I am not sure I would trust a free program of this size. Who knows what type of bloatware, spyware, or other malicious stuff may be imbedded in it.

But then again I am a paranoid old man. If you know your system you should know what it can do. I know my games play maxed out flawless at 2k, and my system is stable at 5.0ghz.

 

[boju]

2.2 gb is a lot for a benchmarking tool I have never heard of. Quite frankly. I am not sure I would trust a free program of this size. Who knows what type of bloatware, spyware, or other malicious stuff may be imbedded in it.

But then again I am a paranoid old man. If you know your system you should know what it can do. I know my games play maxed out flawless at 2k, and my system is stable at 5.0ghz.

3DMark basic is 3.8GB Think this mob might be the new free 3DMark

 

[Dosflores]

2.2 gb is a lot for a benchmarking tool I have never heard of. Quite frankly. I am not sure I would trust a free program of this size. Who knows what type of bloatware, spyware, or other malicious stuff may be imbedded in it.

Malware doesn't need more than a few KB to destroy your system. I'd rather not trust a graphics benchmark too small to contain high-quality graphics assets.

I think visiting tomshardware.com is more dangerous than installing a benchmark tool that you have never heard of, if it seems legit. Malicious ads are everywhere. I have already experienced ads on tomshardware.com that took me out of the site entirely, to try to scam me :lol:

 

[gdmaclew]

I thought Tom's checked ads before including them on their sites.

 

[Ilya__]

The more ways to test the GPU the better!

 

[Eximo]

Well the ad providers are supposed to... I have gotten malware from Tom's once or twice, but also many other legitimate places.

I just want that stupid script on the homepage to die in a fire. Refuses to load on my system so I have to wait about one minute for it to fail before I can use the browser again. And that newsletter pop-up conflates the issue since when it displays it automatically takes you to wherever you clicked, which is not what I intended to do. That means loading back up the home page and waiting for another minute.

The auto-playing videos would be nice to disable as well. I literally have no idea what is in those since I kill them immediately. I've stopped going to some sites that have those sitting off screen restarting themselves every ten seconds. (Also not a fan of infinitely expanding pages, all kinds of bugs and poor performance from that) I clicked on one article, don't need to automatically load the rest when I get towards the bottom. Having to scroll back up to look at the comments is quite annoying.

I get ad revenue, but anything that detracts from the user experience shouldn't be on a website. Relying on accidental clicks is all that drives the industry, I swear. I shouldn't have to install an ad-blocker just to go to websites. Defeats the whole purpose anyway.

Sponsored links seem to be the way to go with tech sites, but they have to be good. Tom's not so much. Whenever I see their price lists for things in the article they are way off. Tom's daily deals should be more regularly updated if they want revenue that way.

 

[johnrhenle]

I ran it. An R9 290 scored 1603. Interestingly, it kept putting CPU details (processor brand, clock speed) on the screen instead of GPU details. And the benchmark wouldn't run a second time, or any of their other tests. Kept telling me to check my internet connection, which is working fine. Loaded it on my phone. It hangs, then wants to download a giant data file, but never succeeds. On their website, if you'r patient enough to wait for it they have some very limited results listed. Most recent phones score higher than 1600. R9 290 is a couple years old now, but there is no way in hell it gets outperformed by the GPU in cell phones.

Looks like an Apples to raisins type of test. No way to actually compare GPU performance from platform to platform. And the software seems to be very buggy on both platforms I looked at.

 

[PdxPetmonster]

I couldn't get the damned thing to run correctly, kept giving me a non-specific error when I did the official and custom tests. The Experience didn't do much of anything either. I'll have to wait for the next revision of this to see if that works.

 

[Giroro]

I'm not saying this mines crypto currency for its creators. But if I were designing a freeware program designed to put 100% load on a GPU, then I would totally make it mine me some altcoins.

 

[alextheblue]

I ran it. An R9 290 scored 1603. <snip> Most recent phones score higher than 1600. R9 290 is a couple years old now, but there is no way in hell it gets outperformed by the GPU in cell phones.

Looks like an Apples to raisins type of test. No way to actually compare GPU performance from platform to platform. And the software seems to be very buggy on both platforms I looked at.

If you run the PC version on default, it runs a PC-specific workload. You can however, choose to run the mobile tests on PC. That's the only way to even begin to get comparable results. They still aren't 100% comparable at this time due to different OS, APIs, etc but it's undoubtedly better than previous tools.

IMO they should compile a native ARM64 version for Windows on ARM, so we can test some QC SoCs against x86 iGPUs on the same platform. But if that happens, it probably won't be until some time after they get the DX12 renderpath up and running.

 

[bit_user]

I ran it. An R9 290 scored 1603.
...
Most recent phones score higher than 1600. R9 290 is a couple years old now, but there is no way in hell it gets outperformed by the GPU in cell phones.

The results shown on their website are for different tests for PC vs. phone. Unfortunately, it doesn't seem to give you the ability to see apples-to-apples comparison between the two.

I think the only PC graphics a high-end phone could reasonably challenge are Intel's integrated GPUs. In fact, I visited their website in hopes of seeing exactly that.
: (

 

[bit_user]

if I were designing a freeware program designed to put 100% load on a GPU, then I would totally make it mine me some altcoins.

People would figure it out pretty quickly, at which point your reputation would be irreparably damaged.

I've uninstalled benchmarking apps that were unexpectedly draining my battery and I'll never put them back. Bad review filed, as well.

 

[Giroro]

if I were designing a freeware program designed to put 100% load on a GPU, then I would totally make it mine me some altcoins.

People would figure it out pretty quickly, at which point your reputation would be irreparably damaged.

I've uninstalled benchmarking apps that were unexpectedly draining my battery and I'll never put them back. Bad review filed, as well.

Would it really damage your reputation though? Are PC users so entitled and vindictive that they would rather spin their wheels doing nothing while running a benchmark instead of supporting the creator of some free software?
If you're running the GPU at full load regardless, why not have it doing some actual work in the process?

Also, why is it surprising that running a benchmark would drain your battery?

 

[R0GG]

The program and website looks pretty legit, although the 2.2 GB could discourage an accidental user from downloading and running it if doesn't give a special advantage over the concurrent well established GPU benchmarking softwares.

 

[dennimonh]

NO support for AMD...obviously this has been modified to run better on intel and nvidia...no fair play

 

[bit_user]

NO support for AMD...obviously this has been modified to run better on intel and nvidia...no fair play

It doesn't say "no support for AMD", it says:

and boasts generous support for various popular graphics APIs thanks to Basemark’s cooperation with several leading companies that include Imagination Technologies, Intel, NVIDIA, Qualcomm, and Renesas. Version 1.0 currently supports OpenGL 4.5, OpenGL ES 3.1, and Vulkan 1.0.

If they support OpenGL and Vulkan, the only way it wouldn't support AMD is if they require some combination of extensions available from all those players but not AMD. That seems unlikely.

Now, if this features Nvidia-specific optimizations, then it should become apparent and that would damage its credibility as a benchmark. So, it's not in their interest to do vendor-specific optimizations. That doesn't mean they aren't, but any such accusations should be evidence-based, rather than blaming them on the basis of association.

 

[bit_user]

if I were designing a freeware program designed to put 100% load on a GPU, then I would totally make it mine me some altcoins.

People would figure it out pretty quickly, at which point your reputation would be irreparably damaged.

I've uninstalled benchmarking apps that were unexpectedly draining my battery and I'll never put them back. Bad review filed, as well.

Would it really damage your reputation though? Are PC users so entitled and vindictive that they would rather spin their wheels doing nothing while running a benchmark instead of supporting the creator of some free software?
If you're running the GPU at full load regardless, why not have it doing some actual work in the process?

It's not about being vindictive - it's about the program not fulfilling its purpose. The goal of a graphics benchmark is to inform hardware purchasing decisions and let people know how well games & VR content will run on their device. It can't do a meaningful amount of crypto-mining without compromising that goal.

Moreover, crypto-mining costs people money in the form of electricity costs and decreased battery life. So, it's not a free lunch.

Finally, yes, if I download and run a program claiming to be a benchmark, but it's actually a distributed crypto-mining vehicle, that's misrepresentation and basically qualifies it as a trojan horse. That definitely warrants negative reviews and potentially removal from various app stores.

Nobody told them it had to be free. If I had the option to pay a few $ for it, I would. Making it free doesn't legitimize having it crypto-mine without disclosing that fact.

Also, why is it surprising that running a benchmark would drain your battery?

more than necessary to accomplish its stated purpose, I mean.