Firing Squad Benchmarks Redone

[shogun]

Thresh's firing squad has redone their quake 3 benchmarks at <a href="http://www.firingsquad.com/hardware/q3videoroundup2/default.asp">http://www.firingsquad.com/hardware/q3videoroundup2/default.asp</a> using the TurboGL drivers finally. Check it out, the results are more realistic now with the G400MAX getting much higher scores.

[alessandro]

....but if you check the summary of the test , again the G400MAX increase is not mentioned at all, even if it wins the TNT2 in all tests (They really must HATE our card)!!!!!!!

[Helevitia]

I just don't understand what FiringSquad was trying to accomplish? First of all, they are comparing apples to oranges. You can't compare a 4th gen vid card to a 3rd gen vid card. They keep talking about how the DDR cards did so well over the rest and that's it! They do mention how crappy the Maxx is in lower scores but they never metnion the improved MAX scores or the TNT2U for that matter. I always liked FiringSquad, what happened?

Dave

------------------
I can never think of a good signature...~

[paulcs]

Hmmmm. Since when is TurboGL an "unofficial" or "unconventional" driver. Looks like the original goal, to undo a mistake my redoing the Max tests, was intentionally obscured by this "unofficial" driver business.

I think this particular series was poorly executed from the get go. At least the numbers look reasonably accurate this time around.

Paul
paulcs@flashcom.net

[Himself]

I think the generation thing is a bit overdone, as long as they stick to the same price range I think it's ok.

[DuRaNgO[MU]]

I was at a lan party this weekend, there were ~12 people, 5-6 had voodoo's of some sort, 5-6 had tnt's of some sort, then there was me , with my souped up g400, everyone commented on the visuals, usually with there mouth hanging open, I don't really care if other people believe the crap on the web, "I like to be different, and better" :P
jim


------------------
Abit BE6 with P3 500 @ 5*124=620
14.6 gb Maxtor ata66 and 128mb ram
G400 vanilla 32mb @ 168/210
Sblive with Altec Lansing speaker combo
384k DSL and Realtek nic
Windows 98se with DX7a
3dMark2000 = 2956
Worn out reset button :O)

[ALBPM]

Heheh... Don't you just love that DuRaNgO!!

I saw the same thing happen at a LAN Party back in October. It's funny how no one badmouths Matrox once they can see and compare the Quality of the graphics against VooDoo, nVidia and ATI.
Besides these people were getting down right religious as they looked over Kruzin's, Rag's and my shoulders...Oh Gawd!!, Jesus!! Gawd Damn Look at That!!...

Paul

[treilanin]

I tend not to put too much faith in whatever Thresh's Firing Squad. Their articles seem to be pretty thin when it comes to reviews and quite frankly the reviews done lack any depth. Like I care if I a bagillion frames per second on an NVidia card and only a bagillion - 5 frames per second on the G400. I have thoroughly been happy with my G400 and will be so for quite some time.

[Darin]

Their QIII normal results for the Cely 450 are VERY close to my cely 490 numbers @ 1024x768. Unfortunately, their PIII 7something scores didn't really improve at this resolution, which sugests that this is a fillrate limitation. I was planning on upgrading my processor in the next month, and was was hoping for improved performance, but now I'm not so sure. Are any of you getting much higher scores @ 1024x768 with faster processors?

[Sensei]

All these tests show is that a 4th gen card is better than a 3rd gen card. It also shows that the 3rd gen cards ended up at about the same speed when all was said and done. I guess the similar fillrates finally kinda evened the cards out after all.

Sen

[franksch3]

Darin,

Quake 3 is mostly fillrate limited at 1024x768, depending on your settings. A faster processor doesn't quite help.

The reason why the GeForce scores so high is more because of it's higher fillrate than using T&L in Quake 3.

One little note about the rants about a month ago about a PIII 600+ processor scoring better in software T&L than hardware T&L on the GeForce. It could very well be true that it scores higher in software T&L.

However, many sites forget that during a benchmark no other things are done, like AI, gravity calculations, collision detection, etc. These things are also quite CPU intensive, but are not used when running a timedemo. Because of this the real fps in the game is probably higher with hardware T&L, because T&L and those other calculations are done in parallel.

[Corwin_Brute]

To Darin :
In high resolutions, the benchmark is not CPU-limited, but fill-rate limited. So I'm not sure you would get a huge increase by upgrading your CPU. Of course, a faster CPU is always welcome...

As for comparing 3rd generation card and 4th generation card, I'm ok, as long as they are in the same price range and offer at least similar features.

About FS benchmark : it was aimed at hardcore Q3 players, and these people are better off with a GeForce, which is what the benchmark shows. But they should have included TGL in the first place...

------------------
Corwin the Brute

[JC]

I think its great that 4th generation cards are reviewed along side 3rd generation. I think that some people would like to believe their 3rd gen. card is still the fastest.

[Ozymandis]

The funny thing none of the reviewers EVER mention is that in HQ mode the TNT2 Ultra doesn't do trilinear filtering, so g400 gets a massive hit for it and the results are closer than they should be. On my TNT there is NO performance hit for trilinear, and on my g400 at 1024x768 all options highest there is a 33% hit for trilinear.

------------------
p3-500, 128mb, g400max, wd hd, promise, 3com ethernet connected to our school's 400mbits of bandwidth...

[Marcelo]

33% performance hit when enabling trilinear filtering sounds about right.