Catalyst 10.3 Benchmarks @WSGF

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Apocalypse23

Golden Member
Jul 14, 2003
1,467
1
0
Here are my benchmark results after installing 10.3 beta, from 10.2:

Asus 5850 ocd to 825MHz core and 1025Mhz memory
E8400 @ 4.0Ghz
CCC settings:
AA- application control
AF- application control
Catalyst AI- Standard
Mipmap detail- High Qaulity
Wait for vertical refresh: Off/Application control
AA Mode - Super Sample AA
Open GL Settings - unchecked triple buffer




Furmark performance@ 1440x900, rest was on default:
10.2 driver:
FPS Min=74 Max=150 Avg=98

10.3 beta driver:
FPS Min=77 Max=153 Avg=100

3d Mark Vantage performance:
10.2 driver:
P12548
GPU score: 15602
CPU score: 7906

10.3 beta driver:
P12787
GPU score: 16117
CPU score: 7893

So there is some general improvement with these drivers.
 
Last edited:

Kuzi

Senior member
Sep 16, 2007
572
0
0
I'm not sure if these numbers are true, but these are the 5870 numbers from the graph:

Batman:
1680x1050:
9.11: 77
10.3: 115
Gain: 49%

1920x1200:
9.11: 65
10.3: 95
Gain: 46%

Dirt 2:
1680x1050:
9.11: 54
10.3: 74
Gain: 37%

1920x1200:
9.11: 50
10.3: 68
Gain: 36%

HAWX:
1680x1050:
9.11: 83
10.3: 96
Gain: 15%

1920x1200:
9.11: 78
10.3: 88
Gain: 12%

STALKER:
1680x1050:
9.11: 31
10.3: 43
Gain: 38%

1920x1200:
9.11: 29
10.3: 34
Gain: 17%
 
Nov 26, 2005
15,086
310
126
:)

I actually have 2 Vista U 64 copies. But Unreal Tournament III plays much better (faster) on XP Pro; I've tried both, hehe
 

Mem

Lifer
Apr 23, 2000
21,476
13
81
Last edited:

lifeblood

Senior member
Oct 17, 2001
999
88
91
When are the 10.3 due to be released?

They need to release them far enough in advance of the Fermi release to insure they are the drivers used for any benchmarks. ATI will want to put their best foot forward for that.
 

konakona

Diamond Member
May 6, 2004
6,285
1
0
http://www.neowin.net/news/market-share-windows-7-up-os-x-down-chrome-up-ie-down

Windows platform gained from 92% to 92.12% percent while OS X fell from 5.16% to 5.02%. Windows 7 made a gain from the previous month growing from 7.57% to 8.92%, XP dropped from 66.15% to 65.49%, while Vista also dropped from 17.47% to 16.51%.

65%!

yeah, but 'most people' use PoS computers with an IGP or something equivalent. I am pretty sure he meant us enthusiasts by 'people'. c'mon, a computer with 4gb ram deserves better and you know that ;)
 

RussianSensation

Elite Member
Sep 5, 2003
19,458
765
126
yeah, but 'most people' use PoS computers with an IGP or something equivalent. I am pretty sure he meant us enthusiasts by 'people'. c'mon, a computer with 4gb ram deserves better and you know that ;)

Yes, and most people also care about value.

Paying $250 before taxes for an 'upgrade' version of Windows 7 Professional is ridiculous for most people: http://www.futureshop.ca/en-CA/produ...6ed41f645cen02

Apple charges what $30 for an upgrade: http://store.apple.com/us/product/MAC_OS_X_SNGL

Windows 7 is just an evolution of Vista so it should also cost $30-40 max.

I have used both Windows XP and Windows 7. Do I think Windows 7 is better than Windows XP? Yes. Do I think it's worth $285 CDN for an upgrade version? No. It's one thing if you are buying a new computer, of course you would choose Windows 7 over XP since it's included in the price already.

BTW, I purchased Windows 7 Pro Upgrade for $39.99 CDN as a student. But I am not sure everyone else is eager to spend $249.99 for the same thing! Even Home Premium is $129.99.
 
Last edited:

konakona

Diamond Member
May 6, 2004
6,285
1
0
I guess I am a bit spoiled, took that $30 student upgrade option for granted. Got both of my copies free thru MSDNAA. Man, didn't know it was THAT expensive though :eek: $250 for an OS is hella expensive. I totally wouldn't mind going back to vista, the two aren't that different in my usage.
 

Rezist

Senior member
Jun 20, 2009
726
0
71
They need to release them far enough in advance of the Fermi release to insure they are the drivers used for any benchmarks. ATI will want to put their best foot forward for that.

Never thought of that but yeah no wonder.
 

zerocool84

Lifer
Nov 11, 2004
36,041
472
126
Yes, and most people also care about value.

Paying $250 before taxes for an 'upgrade' version of Windows 7 Professional is ridiculous for most people: http://www.futureshop.ca/en-CA/produ...6ed41f645cen02

Apple charges what $30 for an upgrade: http://store.apple.com/us/product/MAC_OS_X_SNGL

Windows 7 is just an evolution of Vista so it should also cost $30-40 max.

I have used both Windows XP and Windows 7. Do I think Windows 7 is better than Windows XP? Yes. Do I think it's worth $285 CDN for an upgrade version? No. It's one thing if you are buying a new computer, of course you would choose Windows 7 over XP since it's included in the price already.

BTW, I purchased Windows 7 Pro Upgrade for $39.99 CDN as a student. But I am not sure everyone else is eager to spend $249.99 for the same thing! Even Home Premium is $129.99.

And you have to pay that amount EVERY time they release a new version. I guarantee you many people have paid more for OSX than they would pay for a $100 upgrade version of windows. Also I don't know why you're even comparing Pro. Regular users are NOT going to use anything that Pro has to offer.
 

vj8usa

Senior member
Dec 19, 2005
975
0
0
Yes, and most people also care about value.

Paying $250 before taxes for an 'upgrade' version of Windows 7 Professional is ridiculous for most people: http://www.futureshop.ca/en-CA/produ...6ed41f645cen02

Apple charges what $30 for an upgrade: http://store.apple.com/us/product/MAC_OS_X_SNGL

Windows 7 is just an evolution of Vista so it should also cost $30-40 max.

How about actually linking to a fair comparison?
http://www.newegg.com/Product/Produc...82E16832116713
Home Premium upgrade, $110.

"Most people" aren't going to need what Professional offers over Home Premium. Also, even Professional is only $175 at Newegg (US). Not sure why you're comparing inflated Canadian prices for Windows to American prices for OS X.

Oh, and as per Wikipedia:
"This version of Mac OS X [10.6] focuses on improving performance, efficiency and reducing its overall memory footprint compared with its predecessor Mac OS X v10.5 "Leopard", rather than new end-user features."

So you're essentially paying $30 for the kind of stuff you'd get for free in a Windows service pack? Can't say I see the value here.
 
Last edited:

BD231

Lifer
Feb 26, 2001
10,568
138
106
This is quite astonishing, I never thought drivers would matter that much

Just an FYI, new, high end video cards almost *always* see performance improvements via driver updates early in their lifespan; sometimes even later. nVidia and ATI alike.
 

1h4x4s3x

Senior member
Mar 5, 2010
287
0
76
Just an FYI, new, high end video cards almost *always* see performance improvements via driver updates early in their lifespan; sometimes even later. nVidia and ATI alike.
Certainly, but not to such an extent.

Amazing. Like I said in the other thread, I've been doing this automated benchmarking for 10 years with different games. Percentage wise, this is the biggest increase I've ever seen from one driver release for any game over that period of time. This leads me to believe one of two things:

1> Something is wrong with the driver and it isn't rendering everything correctly in DX11. I don't have any evidence of this, and obviously I hope it's not the case.
2> There were major performance gains to be had on the DX11 path. This would bode well for BFBC2.
http://www.widescreengamingforum.com/forum/viewtopic.php?f=64&t=18459&start=30

My 3) They were sandbagging :sneaky:
 

toyota

Lifer
Apr 15, 2001
12,957
1
0
I guess I am a bit spoiled, took that $30 student upgrade option for granted. Got both of my copies free thru MSDNAA. Man, didn't know it was THAT expensive though :eek: $250 for an OS is hella expensive. I totally wouldn't mind going back to vista, the two aren't that different in my usage.
why does everyone keep saying upgrade? they had full retail boxed versions of Windows 7 Pro for $29.99 for several months. all you needed was an .edu email address.
 

toyota

Lifer
Apr 15, 2001
12,957
1
0
This is quite astonishing, I never thought drivers would matter that much.
Though, as he said, older titles saw little to no improvement.
yeah its very rare that it happens. I know when I first got my 8600gt it was no better than a 7600gt in some games with release drivers but then the next two driver releases doubled the framerate in many games.
 

Keysplayr

Elite Member
Jan 16, 2003
21,209
50
91
Are the 10.3's official yet? I went to AMDs site and still only 10.2 is the latest I can download.
 
Dec 30, 2004
12,554
2
76
If those numbers are true, I'm quite suprised. I didn't think drivers could really make that kind of improvement... unless AMD was sandbagging or there were bugs that they worked out to get the performance. I'm looking forward to these drivers now. :)

Maybe Catalyst 10.3 is AMD's answer to Fermi...

I think they were sandbagging.
 

Keysplayr

Elite Member
Jan 16, 2003
21,209
50
91
nope. thats why some people are asking if they should try them or just stick to the 10.2 for now.

Ah, ok. I hope they're made official soon. I'd like to use the 10.3s for my benches. No point in doing the benches right now with 10.2's if the 10.3's offer more performance and right around the corner. I guess I could use the beta, but I am reluctant to use it due to public perceptions of the results. They would be received with mixed feelings I suppose.
 

waffleironhead

Diamond Member
Aug 10, 2005
6,917
429
136
They need to release them far enough in advance of the Fermi release to insure they are the drivers used for any benchmarks. ATI will want to put their best foot forward for that.

I wonder how many sites will re-benchmark versus using the data they already collected. I'm guessing they will be spending enough time running fermi thru its paces and just rely on existing data for cypress.
 

Lonyo

Lifer
Aug 10, 2002
21,939
6
81
I wonder how many sites will re-benchmark versus using the data they already collected. I'm guessing they will be spending enough time running fermi thru its paces and just rely on existing data for cypress.

One would hope that they would re-run Cypress benchmarks slightly in advance of Fermi being available (depending on how the timings work).
Either they should be able to run both, or they should be able to re-run Cypress if Fermi is a late arrival (e.g. they get cards only a couple of days before NDA lifts).

Certainly assuming these results hold it will be necessary to check drivers before looking at reviews. I think AT at least usually runs updated numbers though, for the most part. Hopefully they will if the 10.3s are out in time.