Does the 2900xt really blow this badly?

[Truckchase!]

See benchmark results 1900xtx vs. 2900xt. Ran drivercleaner.net between installs and cleared out all performance info. I can't seem to figure out why this card seems so slow. Mind you, the 1900 is overclocked and this is at stock 3d clocks, but I still figured the 2900 would beat it handily....

1900:

Benchmark Log File created on 8/14/2007 at 6:39:14 PM

Number of cores 2
Specification Intel(R) Core(TM)2 CPU 6600 @ 2.40GHz
Core Stepping B2
Core Speed 3488.2 MHz (8.0 x 436.0 MHz)
Stock frequency 2400 MHz
L2 cache 4096 KBytes, 16-way set associative, 64-byte line size

Northbridge Intel P35/G33/G31 rev. A2

Memory Size 4096 MBytes
Memory Frequency 545.0 MHz (4:5)
CAS# 5.0
RAS# to CAS# 5
RAS# Precharge 5
Cycle Time (tRAS) 15

DMI BIOS
date 05/25/2007
Motherboard Info
vendor ASUSTeK Computer INC.
model P5K Deluxe
Windows Version Microsoft Windows Vista (6.0) Business Edition (Build 6000)
DirectX Version 10.0

Video:ATI @ x1900xtx/729 1818 Catalyst 7.8
Audio:Sound Blaster X-fi Fatal1ty

Tests performed at 1600x1200x32, 4x FSAA & 16x Anistropic (Quality setting)
************************************************** ******************************

Results for Quake2:
Result 1:371.0
Result 2:370.9
Result 3:371.0
Result 4:371.1
AVERAGE FPS: 371
***************************************

Results for Quake3:
Result 1:353.6
Result 2:342.3
Result 3:353.3
Result 4:353.3
AVERAGE FPS: 350.62
***************************************

Results for UT2004FlyBy:
Result 1:274.3
Result 2:274.3
Result 3:274.3
Result 4:274.3
AVERAGE FPS: 274.3
***************************************

Results for UT2004Botmatch:
Result 1:74.73
Result 2:74.89
Result 3:74.77
AVERAGE FPS: 74.8
***************************************

Results for Comanche4:
Result 1: 98.1
Result 2:100.4
Result 3:100.0
AVERAGE FPS: 99.5
***************************************

Results for FarCry:
Result 1:102.7
Result 2:102.5
Result 3:102.5
AVERAGE FPS: 102.57
***************************************

Results for HalfLife2:
Result 1:146.2
Result 2:146.2
Result 3:146.2
AVERAGE FPS: 146.2
***************************************

Results for SeriousSam2:
Result 1:101.2
Result 2:101.0
Result 3:101.2
AVERAGE FPS: 101.13
***************************************

Results for Doom3:
Result 1:87.1
Result 2:87.2
Result 3:87.2
AVERAGE FPS: 87.17
***************************************

Results for HalfLife2_Lost_Coast:
Result 1: 62.3
Result 2: 62.3
Result 3: 62.3
AVERAGE FPS: 62.3
***************************************

Results for Quake4:
Result 1:64.11
Result 2:63.88
AVERAGE FPS: 64
***************************************

Results for SplinterCell_Chaos_Theory: (Note that AA settings do not apply to Splinter Cell + HDR)
Result 1:75.95
Result 2:75.93
Result 3:75.93
AVERAGE FPS: 75.94
***************************************

Results for Supreme_Commander:
Result 1:28.297
Result 2:27.963
AVERAGE FPS: 28.13
***************************************

Results for Company_of_Heroes_DX9:
Result 1:73.43
AVERAGE FPS: 73.43
***************************************

Benchmarks finished at 7:39:37 PM

***************************************
GRINGOMARKtm=59
***************************************

2900:

Benchmark Log File created on 8/15/2007 at 12:17:34 AM

Number of cores 2
Specification Intel(R) Core(TM)2 CPU 6600 @ 2.40GHz
Core Stepping B2
Core Speed 3488.1 MHz (8.0 x 436.0 MHz)
Stock frequency 2400 MHz
L2 cache 4096 KBytes, 16-way set associative, 64-byte line size

Northbridge Intel P35/G33/G31 rev. A2

Memory Size 4096 MBytes
Memory Frequency 545.0 MHz (4:5)
CAS# 5.0
RAS# to CAS# 5
RAS# Precharge 5
Cycle Time (tRAS) 15

DMI BIOS
date 05/25/2007
Motherboard Info
vendor ASUSTeK Computer INC.
model P5K Deluxe
Windows Version Microsoft Windows Vista (6.0) Business Edition (Build 6000)
DirectX Version 10.0

Video:ATI @ x2900xt 512MB 743/1656 Catalyst 7.8
Audio:Sound Blaster X-fi Fatal1ty

Tests performed at 1600x1200x32, 4x FSAA & 16x Anistropic (Quality setting)
************************************************** ******************************

Results for Quake2:
Result 1:415.0
Result 2:415.3
Result 3:411.6
Result 4:420.7
AVERAGE FPS: 415.65
***************************************

Results for Quake3:
Result 1:316.3
Result 2:316.3
Result 3:316.1
Result 4:316.3
AVERAGE FPS: 316.25
***************************************

Results for UT2004FlyBy:
Result 1:224.4
Result 2:224.5
Result 3:224.5
Result 4:224.5
AVERAGE FPS: 224.48
***************************************

Results for UT2004Botmatch:
Result 1:69.57
Result 2:70.12
Result 3:69.85
AVERAGE FPS: 69.85
***************************************

Results for Comanche4:
Result 1:105.1
Result 2:105.9
Result 3:106.0
AVERAGE FPS: 105.67
***************************************

Results for FarCry:
Result 1:81.81
Result 2:81.72
Result 3:81.72
AVERAGE FPS: 81.75
***************************************

Results for HalfLife2:
Result 1:112.7
Result 2:112.8
Result 3:112.8
AVERAGE FPS: 112.77
***************************************

Results for SeriousSam2:
Result 1:78.2
Result 2:77.9
Result 3:77.8
AVERAGE FPS: 77.97
***************************************

Results for Doom3:
Result 1:114.5
Result 2:114.6
Result 3:114.5
AVERAGE FPS: 114.53
***************************************

Results for HalfLife2_Lost_Coast:
Result 1: 65.2
Result 2: 65.4
Result 3: 65.4
AVERAGE FPS: 65.33
***************************************

Results for Quake4:
Result 1:88.64
Result 2:81.67
AVERAGE FPS: 85.16
***************************************

Results for SplinterCell_Chaos_Theory: (Note that AA settings do not apply to Splinter Cell + HDR)
Result 1:100.4
Result 2:100.4
Result 3:100.4
AVERAGE FPS: 100.4
***************************************

Results for Supreme_Commander:
Result 1:30.502
Result 2:30.311
AVERAGE FPS: 30.41
***************************************

Results for Company_of_Heroes_DX9:
Result 1:76.62
AVERAGE FPS: 76.62
***************************************

Benchmarks finished at 1:19:15 AM

***************************************
GRINGOMARKtm=59
***************************************

[NKD]

well hd2900 doesnt really suck, well in some games there are still driver issues where its even slower than x1900, but as i see in most of the games its doing quite good, and u wont see that much of performance advantage in some of the games unless u bump up the resolutiong to 1920x1200. the drivers are really improving though, i believe in quake 2 or the older games u wont see any difference what so ever.

[adamsleath]

mmmmmakes that oced x1900xtx look good.

[zanzabar]

i bet that if u ran 2x aa the 2900 would have a much larger margin, and what 2900 are u running the ddr3 card dont have the memory bandwidth to make up for bad texture filtering drivers like the ddr4 dose, adn by useign 16x af ur running 8x af and then another 16x over it since software dx10 forces 8xaf on top of what the software dose

[Truckchase!]

Here's some images for easier reading....

MAN, this is disappointing.

Edit: These are @ 1600x1200, 4xAA, 16x Aniso

[purecain]

redo these tests with the latest drivers.... these test are no longer valid....

[Truckchase!]

That's cat 7.8 on both cards...

[DilTech]

redo these tests with the latest drivers.... these test are no longer valid....

That is with 7.8.

[NKD]

i think something is seriously wrong with those results, i know there are some game 2900xt loses to 1900xt but all the review sites say its a driver glitch, but i havent seen it loose to that many games anywhere in any review.

[Truckchase!]

i think something is seriously wrong with those results, i know there are some game 2900xt loses to 1900xt but all the review sites say its a driver glitch, but i havent seen it loose to that many games anywhere in any review.

I sure hope you're right, because I've wasted almost 400 bux otherwise. Unfortunately as of right now I can't find anything wrong with the tests. I'm currently running another set without AA to see what the results are, and I'm also investigating the possibility that something is wrong with the config files for each game after a gfx card change. I've never had to swap out cfg files for any of the games negatively impacted in the past, however, when I do an upgrade. My initial thorough evaluation of Farcry has resulted in nothing but completely reproducible, crappy results.

Mind you again, the x1900xtx was overclocked to 729/1818 and the 2900xt is at 743/1656(stock 3d) but I can't believe the 2900 can still get trounced so handily in Halflife 2, Farcry (which is the regulator map test, btw), and Serious Sam 2. (not to mention that even when it doesn't get trounced it isn't much faster in most cases)

[purecain]

i dont get it, you say you had crappy results with farcry.... my cards flew single and crossfire..... results were in eocf forums....

[ChaosMinionX]

i dont get it, you say you had crappy results with farcry.... my cards flew single and crossfire..... results were in eocf forums....

Could you post some of your own results? You are losing credibility when you just say "my cards flew" and the "these tests are no longer valid" statements due to driver versions.

Yeah I could see you defending the HD2900XT, by all means I want it to be a great card in driver revisions, but dont make the rest of the red side look bad by throwing out FUD and saying your results are different. Actually show them, and
make a good case for the R600

Just a lil input from what I have seen you post here, and havent really given anything productive beyond that cat7.8's were a nice improvement to the HD2900XT

[zanzabar]

is that the 512 or the 1gb 2900

[Truckchase!]

512mb.

[adamsleath]

i see the results - very clearly ;)
what are the stock speeds of your x1900xtx? [650/1550??]

[Truckchase!]

i see the results - very clearly ;)
what are the stock speeds of your x1900xtx? [650/1550??]

That's the only "ray of hope" of sorts; My ultimate intent is to compare OC vs. OC speeds since that's all I play at anyhow, but I haven't gotten around to clocking the 2900xt yet. These tests are with the x1900xtx@ 729/1818(oc) and the 2900xt@ 743/1656.(stock) I still expected the 2900 to win relatively handily however. :-/

I'll post non-AA results tonight and start working on overclocking.

[SNiiPE_DoGG]

with cat 7.8 and 2900xt 1gb 853/1050 i can run call of juarez on max settings and 4x aa @ 1920x1200 res

fps never drops below maybe 40 (no exact test, but pretty good)

maybe you have a problem in your system? maybe its vista buisness? (i run ultimate)

[DilTech]

with cat 7.8 and 2900xt 1gb 853/1050 i can run call of juarez on max settings and 4x aa @ 1920x1200 res

fps never drops below maybe 40 (no exact test, but pretty good)

maybe you have a problem in your system? maybe its vista buisness? (i run ultimate)

Screenshot of the number at the end of the time-demo run? :D

Reason I ask, is because with 7.7, you're talking about an average about 1/3rd of what you're saying your minimum is. Even Crossfire had an average less than what you're saying your minimum is.

[zanzabar]

the 512 mb card has slow memory the ddr3 is way slower than the ddr4, just in latency the ddr4 allows a faster speed with a lowwer latency multiplyer, ddr3 is 2/3 while 4 is 4/4, so u get faster speeds with 1/3 more bandwidth

[ChaosMinionX]

the 512 mb card has slow memory the ddr3 is way slower than the ddr4, just in latency the ddr4 allows a faster speed with a lowwer latency multiplyer, ddr3 is 2/3 while 4 is 4/4, so u get faster speeds with 1/3 more bandwidth

Theoretically that works out correct, but in realworld application it doesnt seem so... the GDDR4 would rock if it were BC06 .6ns rated at like 1600mhz, instead of the BC09 .91ns rated at 1100mhz. But if you OC the BC09's to like 1200mhz ytou start to see it pull away significantly from the GDDR3 stuff.

Although I heard the BC06 are going to be used in the upcoming cards.... who knows.

[DilTech]

the 512 mb card has slow memory the ddr3 is way slower than the ddr4, just in latency the ddr4 allows a faster speed with a lowwer latency multiplyer, ddr3 is 2/3 while 4 is 4/4, so u get faster speeds with 1/3 more bandwidth

The thing is, the hd2900xt isn't bandwidth limited. Kind of like feeding massive bandwidth the the A64, it isn't making too big of a difference.

[ColonelCain]

The thing is, the hd2900xt isn't bandwidth limited. Kind of like feeding massive bandwidth the the A64, it isn't making too big of a difference.

Exactly. The R600 almost has too much bandwidth....

[NKD]

yea there is no difference between 1gb and 512mb, r600 already has monster bandwidth and thats where the one gig card looses all the advantage, i have seen three reviews of the 1gb card and it failed to impress me, i have two radeon 2900xt's on the way and with in a week, i should have results for u, to match the halflife and farcry since i have them on my system.

i just have one question for u, did u uninstall all the nvidia files with driver cleaner, u said u had a swap with the nvidia card, i would clean anything nvidia related, and i am sure u have all the games updated with the latest patch if not it could make a lot of diffrence as the drivers are coded for the latest patches of the games. i dont wanna tell u do a clean install because if i had too much on a system i wouldnt do it myself, uless i had another hard drive that i could put windows on and test it out.

[Truckchase!]

Before we get too OT with the 512/1GB deal, here are some additional performance numbers. I ran tests to determine what the hit is from Aniso and AA. It appears to me that the big downfall of this card is still AA. (even after the 7.8 drivers)

First chart is x1900xtx vs. 2900xt with NO AA and NO ANISO. Notice the performance is a bit better comparatively (1900 vs. 2900), although still not where it should be on some titles. EDIT: (see next post, didn't attach correctly.)

Second Chart is x1900xtx performance hit analysis for 16x aniso and 4xAA/16x aniso. I don't have data for just 4xAA. The % illustrated is % of FPS vs. no AA/Aniso, I.E. higher is better, and indicates less of a performance hit.

The third chart is the same, but this time with the 2900xt. Notice how much more adding AA affects the 2900 than it does the 1900.

Mind you, I'm not saying AA is the only problem. The performance numbers for games like Half Life 2 are still WAY low (in my opinion) without any AA or ANISO vs. the last generation. The performance hits are definitely worth noting however.

[Truckchase!]

Missed posting this one. Oops.

[Zytek_Fan]

Some games seem to not be able to utilise the 2900 to its full potential...

[binormalkilla]

The thing is, the hd2900xt isn't bandwidth limited. Kind of like feeding massive bandwidth the the A64, it isn't making too big of a difference.
Yep. Overclocking the memory past my stock 1000 doesn't get me any FPS in Lost Coast stress test, but you get about 100 in 3dmark06. I don't care that much about benching, so I just leave it at stock.
I actually lost ~1 FPS when I tested a 40 MHz OC on the memory compared to stock...left the GPU at stock.
Definitely GPU bottlenecked.

[binormalkilla]

yea there is no difference between 1gb and 512mb, r600 already has monster bandwidth and thats where the one gig card looses all the advantage, i have seen three reviews of the 1gb card and it failed to impress me, i have two radeon 2900xt's on the way and with in a week, i should have results for u, to match the halflife and farcry since i have them on my system.

i just have one question for u, did u uninstall all the nvidia files with driver cleaner, u said u had a swap with the nvidia card, i would clean anything nvidia related, and i am sure u have all the games updated with the latest patch if not it could make a lot of diffrence as the drivers are coded for the latest patches of the games. i dont wanna tell u do a clean install because if i had too much on a system i wouldnt do it myself, uless i had another hard drive that i could put windows on and test it out.
Nope, not true. I am running Rivatuner hardware monitoring to show the GDDR usage, and I have it going to an LCD studio applet for my G15. In BF2142 it used 536 Mb of RAM on Minsk. I am running 16X af, 2X narrow tent aa (so that's 4x samples), maxed in game settings, and 1920*1080.
I can only imagine how much Crysis and Bioshock will use with these settings. I have the Bioshock demo DLing now, so we'll know soon.

[NKD]

Nope, not true. I am running Rivatuner hardware monitoring to show the GDDR usage, and I have it going to an LCD studio applet for my G15. In BF2142 it used 536 Mb of RAM on Minsk. I am running 16X af, 2X narrow tent aa (so that's 4x samples), maxed in game settings, and 1920*1080.
I can only imagine how much Crysis and Bioshock will use with these settings. I have the Bioshock demo DLing now, so we'll know soon.

well ur right more ram does help but i was only talking about performance wise, there is a mere difference between 1gb and 512mb, i dont think its worth the price premium unless there was a 25-35% performance bump, i actually got two hd 2900's 512mb for 600bucks from someone and i am pretty happy i will have enough ram to run in crossfire mode,

[binormalkilla]

Yea I see what you mean, but won't the GPU take a larger performance hit if the memory can't store all of the data? I could be wrong....I haven't seen too many bandwidth tests when compared to the GPU being OCed on the 1Gb card....TBH I haven't even bothered to OC my memory except for 3dmark. Doesn't even show a difference in gaming, as the GPU is the bottleneck. I guess if you OC the GPU far enough it may require some more bandwidth.

[Truckchase!]

Here's numbers for thought. This is with a 512 MB card. The graphs are in the following order:

Mem only overclocked
GPU only overclocked
GPU and MEM overclocked

The overclock speeds for the gpu and mem, respectively, are 858/1796 (the max the CCC will allow, nearly a 15% overclock on both). Results illustrate a heavy bottleneck on the GPU vs. the memory, as was voiced by others earlier in this thread.

EDIT: Sorry, tests are @ 1600x1200 with 4x BOX AA and 16x Aniso enabled. (Vista 32bit) My system details are in the sig below, with the obvious exception of the vid card.

[binormalkilla]

Interesting results....does the data on these bar graphs represent normalized FPS figures? I. E. 100+[(OCFPS-STOCKFPS)/100]? I'm unsure as to what '% no of OC' means!
Great job, I like your style. Straight up data :D

[Truckchase!]

Heh, thanks. Yah... those figures are normalized. Here's the numbers in a spreadsheet if you're interested...

[CrimInalA]

very interesting .
nice charts comparing gpu - mem overclocking performance on the 2900XT Truckchase . :bounces: