graphics card testing methodology

V

ValVed RaY

Limp Gawd
Joined
Jan 10, 2011
Messages
255
No. sometimes they take a while when the subject matter is striking at the way they do things.
 
your not getting much response cause i think its obvious answer or to some degree irrelevant
 
it is irrelevant in most cases, however I have seen previous drivers cause issues even when doing a complete uninstall of previous drivers in addition to using driver clean up utilities.

I also don't think I'm striking their methodology, I'm merely inquiring.
 
What exactly are you asking for? Exactly? and for that matter, why does it matter?

H uses Windows whatever, I would guess partitioned on a Drive with the latest GPU drivers installed, and everything else you or I would use to achieve a fully functional system.

If the drivers don't work well, it is explained in the individual article.

Also, you are looking at apples and oranges.
On the Hardware Canucks article they are reviewing CPU and memory performance variences across two completely different systems, the H review is for a single GPU run on the same MB. The only variable different in the H review is the game.
 
Last edited:
Not only that..Its a safe bet brent or whoever keeps a clean windows image with never any gpu drivers installed before testing out a new card.....but you can always pm kyle since he may not read every thread http://hardforum.com/member.php?u=3 As far as the thread you linked to, I think its very possible he did not use ddu the way myself and others here do.....not to put him down, but these things are never user error right? I see your point but ddu works as good as a fresh image/install for most of us:) In any case i suspect [H] uses a clean image when testing new hardware
 
Last edited:
systems with identical hardware specifications, can have drastic performance deltas.
a way to attempt to minimalize the potential for that can be by using constant methods for benchmarking. are they re-imaging the systems with a baseline image and then installing the necessary GPU drivers? are they simply running the uninstaller and installing the latest drivers? what other software are they using on their systems to capture their data? etc..
 
primetime said:
Not only that..Its a safe bet brent or whoever keeps a clean windows image with never any gpu drivers installed before testing out a new card.....but you can always pm kyle since he may not read every thread http://hardforum.com/member.php?u=3 As far as the thread you linked to, I think its very possible he did not use ddu the way myself and others here do.....not to put him down, but these things are never user error right? I see your point but ddu works as good as a fresh image/install for most of us:) In any case i suspect [H] uses a clean image when testing new hardware
Click to expand...

thank you, I have sent him a pm with the link to this thread. I try to make no assumptions, but you're probably right if I had to guess..
 
ValVed RaY said:
I wasn't trying to infer that HardOcp was affiliated with Canucks. I was merely using their transparent explanation of their setup as an example.

You have answered my question, but you seem a little defensive and angry :rolleyes:
Click to expand...

Hence why I stated in the second post how this usually goes down. They don't take well to questioning how they do their reviews or parallels to other sites. So the response is expected.

And in the case Kyle returns, the question was why your methodology isn't stated in each review as it pertains to the using clean install or update drivers?
 
JustReason said:
And in the case Kyle returns, the question was why your methodology isn't stated in each review as it pertains to the using clean install or update drivers?
Click to expand...


Isnt it though? I havent checked the recent reviews other then the conclusion because modern hardware is boring as fuck. But doesnt every review state in the text fresh install of blah blah with blah blah drivers? At this point I wonder if its not just obvious that its a fresh install. But hell I guess Im just blase about people talking shit about HardOCP, because [H]ard doesnt give two shits what other people think or say.
 
macksomerville said:
Isnt it though? I havent checked the recent reviews other then the conclusion because modern hardware is boring as fuck. But doesnt every review state in the text fresh install of blah blah with blah blah drivers? At this point I wonder if its not just obvious that its a fresh install. But hell I guess Im just blase about people talking shit about HardOCP, because [H]ard doesnt give two shits what other people think or say.
Click to expand...

Thank God! I mean Kyle for exactly that. I like how HardOCP test hardware, also makes it fun trying to poke holes in there methods. I just usually look at a broad range of reviews if I am interested in something and usually HardOCP has the most interesting results or insight at times.

I just wished they used 3dMark to verify that the system is consistent with previous tests and other systems like it and during the same run of tests to make sure something wasn't messed up :p :D
 
noko said:
Thank God! I mean Kyle for exactly that. I like how HardOCP test hardware, also makes it fun trying to poke holes in there methods. I just usually look at a broad range of reviews if I am interested in something and usually HardOCP has the most interesting results or insight at times.

I just wished they used 3dMark to verify that the system is consistent with previous tests and other systems like it and during the same run of tests to make sure something wasn't messed up :p :D
Click to expand...

I don't agree with this. That would be completely irrelevant tbh and a waste of time. They don't give 2 shits about 3dmark because sane people by cards to play games, not jerk off to canned benchmarks. They use the most recent driver available when they test...you cant ask for anything more. And 3dmark can do more harm than good in regard to completing a suicide run achieving "what ever score" means nothing since in most cases the clocks have to much lower testing games the way they do. People will say i always get "what ever" 3dmark score but of course there card was never truly stable, at least not with all the games they fully test.
The way they test is a lot of work as is....I don't blame them for skipping added bullshit;)
 
I didn't see a question in the OP...
ValVed RaY said:
I was wondering what HardOcp's testing methodology is:
Click to expand...
This is a statement of inquiry. This is usually followed up by a question regarding the inquiry, so let us look further.

ValVed RaY said:
here is one for example:
http://www.hardwarecanucks.com/foru...-intel-skylake-ddr3-vs-ddr4-comparison-3.html

This latest review here:
http://hardocp.com/article/2015/12/...directcu_iii_video_card_review/2#.VntFh3lvllc
Click to expand...
Now he provides links to two different sites, comparing apples to oranges. Here's a hint: there is a link in the first paragraph of the [H] review that leads to what you want.

ValVed RaY said:
It doesn't say if this is a clean windows install, or if chipset drivers are installed. etc..
Click to expand...
Now we have detail expounding on the inquiry in the first statement. I personally think it would be irresponsible for a review site to not have an image of a baseline test environment that they always start with. The professionalism that [H] shows makes me feel that it never had to be stated in the first place. But there is still no question here.

ValVed RaY said:
just curious, if a more detailed rundown can be added to the testing methodology.

thank you
Click to expand...
Given the tone of the rest of the post, this sounds like a statement of demand, and the "thank you" at the end adds a condescending edge that the site needs to cater to this particular reader's tastes.

This is the end of the post, and I see no question in sight.
 
Pioneered and stuff!

I didn't find OP condescending at all. While he may not have directly posed a question, the question he was inferring was quite clear. He provided a comparison to a review that stated their testing, and it appeared he was wanting some clarification on how H does their reviews.
People are so quick to jump on anyone that dares to question the Almighty Kyle around here, regardless of intent or tone.

"Can you tell me about xxxx? What about yyyy? Oh, and zzzz! Moar about zzzz! "

"Go away, you're stupid."
"You don't know what you're talking about."
"I pioneered this shit, so fuck off."

Then someone pops in, in defense of the OP, and well, read above.

In regards to 3dMark, I do like seeing these numbers. While not real world, it provides a fantastic metric to compare. "Ok, the games that I play (which often aren't in the reviews) play like this, and at those settings/oc/whatever, I get npoints in 3DM. But they don't use my games, but I'm curious about upgrading. Will it be worth it to me? I know what another 5k points in 3DM looks like, so that gives me some measure of quantifiable delta to compare with."
 
Well the question seemed a bit silly to some of us, because for one it has been asked, and answered before. I remember feeling embarrassed when i asked how they tested the games....I got an answer i didn't expect..He said "We actually play the games, I know its shocking!:D" I was just over complicating it in my mind the hole time.
 
You must log in or register to reply here.
Back
Top