Tom's Hardware.. Nvidia or Ati biased..?

Printable View

Show 100 post(s) from this thread on one page

05-22-2009, 11:09 PM
Rickpatbrown

"Tom's Games" is pretty good though:rolleyes:. That is the whole reason I have a $2000 machine . . . to play flash games!
05-22-2009, 11:16 PM
randomizer

There were valid reasons for Tom's Games getting a facelift (tongue in cheek). It just depends on your point of view. Rob and Ben probably wouldn't call it valid.
05-22-2009, 11:20 PM
SamHughe

My personal experience is GTX 285 performs equally if not slightly faster overall than an oced 4890 in most scenarios. ATI offers a slightly better IQ, cheaper price but much noisier than the 285. I don't believe 4890 cannot pass 260 in Crysis and fall so behind 285.
05-22-2009, 11:38 PM
tajoh111

Quote:

Originally Posted by perkam

Notice something ODD about the game selection?

They didn't test Crysis...where the regular 4890 wipes the floor with the OCed 275:

http://www.xbitlabs.com/images/video...c/crysiswh.png

And here are the 4890 OC numbers:

http://www.xbitlabs.com/images/video...0/crysiswh.png

Perkam

What the point of turning AA on? So you can get unplayable framerates for both cards.

Crysis with a higher resolution and no AA actually yields better image quality(and more importantly playable frame rates), than this test.

In my opinion, unless we are testing multi chip solution that can handle the load, crysis should not be test with AA because its simply an exercise in futility.

This test would be valid if one of the cards yielded playable framerates, but both are equally unplayable.

The only point of turning AA I can imagine is so ATI cards come out looking better than they really are, as xbitlabs is the only website, that has the 4890 coming out ahead a gtx 285 in crysis warhead. The only bloody website on the entire net including anandtech which is probably the 2nd most favorable review for the 4890.

I definitely question the results of this tomshardware review, however I also question xbitlabs review on NV stuff because I have already found lots of biased evidence against them.

My gaming machine has a 4890 and a 4870x2, I don't have any card from the g200 generation.

I wouldn't dare turn on AA and sacrifice the rest of my settings so I can turn on AA.

Crysis warhead is still choppy on this setup and this is with a core i7 at 4.5ghz.
05-23-2009, 01:41 AM
Nedjo

biased? perhaps... but I'd say who ever is testing EndWar is not biased, it's dumb masochist!

http://www.tomshardware.com/reviews/...p,2297-14.html
05-23-2009, 05:02 AM
RaZz!

Quote:

Originally Posted by Nedjo

biased? perhaps... but I'd say who ever is testing EndWar is not biased, it's dumb masochist!

http://www.tomshardware.com/reviews/...p,2297-14.html

quoted from the very same page:

Quote:

This game really isn’t ideal for benchmarking, since a software limiter caps its frame rate at 30 FPS.

yeah, exactly, well... hum... errrr... why use it as a benchmark then? they already say themselves why this game is absolutely useless as a benchmark and yet they still put it in the review. what kind of logic do i need to do smth like that? :p:
05-23-2009, 05:04 AM
randomizer

Well it doesn't indicate bias, it indicates a lack of common sense.
05-23-2009, 05:42 AM
SKYMTL

Or a lack of necessary games on the reviewer's part...
05-23-2009, 05:43 AM
perkam

Quote:

Originally Posted by tajoh111

What the point of turning AA on? So you can get unplayable framerates for both cards.

Crysis with a higher resolution and no AA actually yields better image quality(and more importantly playable frame rates), than this test.

In my opinion, unless we are testing multi chip solution that can handle the load, crysis should not be test with AA because its simply an exercise in futility.

This test would be valid if one of the cards yielded playable framerates, but both are equally unplayable.

The only point of turning AA I can imagine is so ATI cards come out looking better than they really are, as xbitlabs is the only website, that has the 4890 coming out ahead a gtx 285 in crysis warhead. The only bloody website on the entire net including anandtech which is probably the 2nd most favorable review for the 4890.

I definitely question the results of this tomshardware review, however I also question xbitlabs review on NV stuff because I have already found lots of biased evidence against them.

My gaming machine has a 4890 and a 4870x2, I don't have any card from the g200 generation.

I wouldn't dare turn on AA and sacrifice the rest of my settings so I can turn on AA.

Crysis warhead is still choppy on this setup and this is with a core i7 at 4.5ghz.

Stop trying to over-complicate it m8.

The 4890 regular is equal to the GTX 285 in Crysis.

What does that mean? It means that when you go crossfire with the 4890, you get GTX 285 SLI performance for hundreds less.

Perkam
05-23-2009, 05:48 AM
randomizer

Quote:

Originally Posted by SKYMTL

Or a lack of necessary games on the reviewer's part...

That's almost common sense, right?
05-23-2009, 11:25 AM
LordEC911

LoL, nice try OBR...:rolleyes: :down: :shakes:
05-23-2009, 11:44 AM
xdan

Yes Tom's site begin to be useless. I see here that many of you have confidence in xbitlabs, when i used this site for some Cpu comparation , it was told to me by some users in this news topic that xbitlabs are not veridic.
Anyway a good site from my country ROMANIA lead by MONSTRU and his friends, i think you guys know MONSTRU..:up:
http://lab501.ro/wp-content/uploads/...-2560-copy.png
http://lab501.ro/placi-video/nvidia-...-radeon-4890/6
It's an interesant review single CF/SLI, TRI SLI/ 3 4890, STOCK/OC review GTX 275/4890.:up:
05-23-2009, 12:05 PM
ahmad

There are a lot of variables when doing reviews. Unfortunately to this day I haven't seen anyone do it 100% right. IMHO what you need is:

- Exact driver settings for both setups. This is so important because it allows: 1 people to reproduce results, 2 indicates what the game would look like
- Screenshots. I am not talking at every setting, but when comparing ATI vs Nvidia you got performance and IQ. I don't want someone's "it looks better on [...]". Take some uncompressed screenshots and let me decide what looks better. I want to see what I am getting when going for one vs the other. Also reviewer can now compare side by side for our benefit (for those who care)
- System specs.. obviously. I mean really we have no idea what these reviewers are doing, but if we have system specs at least someone can reproduce the results (or try to)

With any one of those pieces missing the review is incomplete.
05-23-2009, 12:14 PM
Eastcoasthandle

Quote:

Originally Posted by SKYMTL

On a regular basis actually and generally ATI cards USED to get better framerates than Nvidia cards in it (DX9 version). However, the last few Nvidia driver releases have improved their performance significantly. 30 fps average is generally playable but in heavily forested areas that 30 can and will be cut significantly and the game will seriously bog down.

If you look at the graph above, it seems more a race to who finishes last rather than any card delivering meaningful performance.

However, looking at the tests on the linked review at Tom's there are some questions that popped up:

- Why test high end cards without 2560 resolution?
- What is their test setup? I see "i7 920" and a bunch of GPUs listed and that's it
- In some cases, AA tests were eliminated at certain resolutions.

All in all the article seems to raise more questions than it answers.

I found Crysis and Crysis Warhead playable at 25 FPS or better. Which is something I cannot say with other games. 30 FPS is certainly not need in order for this game to be playable. But lets not confuse playable and ideal though... :D
05-23-2009, 07:36 PM
RejZoR

4xAF lmao, isn't that stupid by itself? I've never used anything below 16xAF since x1950 Pro. On powerful HD4850 i never even thought of either lowering this setting or even disabling it. Performance hit is negligible, but image quality gain is like off the charts.
Maybe thats why NVIDIA is so much better, because AF hits them so badly?
05-23-2009, 09:55 PM
Stukov

Quote:

Originally Posted by SKYMTL

Actually, some of us still remember the time when Tom's was THE best hardware review website out there. This was before it went downhill with the BoM merger.

Yeah I know, back then, at least for me it wasn Anandtech vs Tomshardware. Even before he sold out it was starting to slip a little bit, they champoined Intel hardware a bit more than they should have (this is back in the A64 vs P4 days).
05-23-2009, 10:33 PM
Newblar

Quote:

Originally Posted by Stukov

Yeah I know, back then, at least for me it wasn Anandtech vs Tomshardware. Even before he sold out it was starting to slip a little bit, they champoined Intel hardware a bit more than they should have (this is back in the A64 vs P4 days).

tom's ph2 review had insane bias in it.. so i dont think the old bias has completely been forgotten
05-23-2009, 10:42 PM
randomizer

I'd like to see someone actually point out bias where it can't be put down to a bad reviewer. Lots of claims, no evidence. I understand that the "THG paid ad" bandwagon does have a lot of room though, so I forgive you for jumping on.
05-23-2009, 11:13 PM
Stukov

Quote:

Originally Posted by randomizer

I'd like to see someone actually point out bias where it can't be put down to a bad reviewer. Lots of claims, no evidence. I understand that the "THG paid ad" bandwagon does have a lot of room though, so I forgive you for jumping on.

It was bias several years ago when they knew what they were doing, now they are just a bunch of monkeys spasming on a football.
05-23-2009, 11:23 PM
randomizer

Well the past is meaningless, nobody goes to read Athlon 64 vs P4 reviews now.
05-23-2009, 11:47 PM
Stukov

Quote:

Originally Posted by randomizer

Well the past is meaningless, nobody goes to read Athlon 64 vs P4 reviews now.

It wasn't meaningless when the present was 4-6 years ago.
05-24-2009, 01:04 AM
randomizer

Quote:

Originally Posted by Stukov

It wasn't meaningless when the present was 4-6 years ago.

Yea... and last I checked 2009 wasn't 4-6 years ago.
05-24-2009, 01:55 AM
Stukov

Quote:

Originally Posted by randomizer

Yea... and last I checked 2009 wasn't 4-6 years ago.

Something is always in the past. You would have to have some pretty good foresight to talk about something occuring in the future or some pretty good timing when talking about something exactly when it happens.
05-24-2009, 05:48 AM
STEvil

Quote:

Originally Posted by randomizer

Yea... and last I checked 2009 wasn't 4-6 years ago.

All the posts previous to this are the past, get out of the thread if you dont like it.
05-24-2009, 07:22 AM
GAR

Quote:

Originally Posted by SKYMTL

Fact: Nvidia destroys ATI in HawX with DX9 and DX10 enabled.

HOWEVER, turn on the DX10.1 option and it is a whole different ball game as ATI wipes the floor with Nvidia.

I wouldnt call 2-5 FPS whipes.

Show 100 post(s) from this thread on one page