3.10 slower than 3.9

3DMark2001 SE ---- 3.9 = 14580 3.10 = 14440

3DMark03 ---------- 3.9 = 4653 3.10 = 4582

9700 Pro - default settings for both drivers
1024*768*32 @ 85 MHz

MOH:AA had some weird texture problems in one map, but I think it was caused
by the mod that was running. Tried 4 other maps and all were OK. COD
worked perfectly, no glitches. Refresh override still works good.

I wish the driver writers would quit worring about FPS and benchmarks and
keep focused on image quality, compatability and stability. What good are a
few extra FPS if the system crashes or displays distorted textures? I guess
it's the benchmarks and FPS that are cited in reviews and that sells cards.

-Kent

"KCB" wrote in message
news:ezuEb.396698$Dw6.1247555@attbi_s02...
3DMark2001 SE ---- 3.9 = 14580 3.10 = 14440

3DMark03 ---------- 3.9 = 4653 3.10 = 4582

9700 Pro - default settings for both drivers
1024*768*32 @ 85 MHz

MOH:AA had some weird texture problems in one map, but I think it was
caused
by the mod that was running. Tried 4 other maps and all were OK. COD
worked perfectly, no glitches. Refresh override still works good.



divide 14580 by 14440. Its not even one percent.
divide 4653 by 4582. Its only 1.5 percent.

i think this is well within the deviation for benchmarks.

You are referring to ~100 points difference. From what I've seen, that is
the margin of error/variaton, for those progs. +/- 100 points, in either,
is negligable.

-
KCB stood up at show-n-tell, in ezuEb.396698$Dw6.1247555@attbi_s02, and
said:

3DMark2001 SE ---- 3.9 = 14580 3.10 = 14440

3DMark03 ---------- 3.9 = 4653 3.10 = 4582

9700 Pro - default settings for both drivers
1024*768*32 @ 85 MHz

MOH:AA had some weird texture problems in one map, but I think it was
caused by the mod that was running. Tried 4 other maps and all were
OK. COD worked perfectly, no glitches. Refresh override still works
good.

--
Strontium

"It's no surprise, to me. I am my own worst enemy. `Cause every
now, and then, I kick the livin' **** `outta me." - Lit

"Strontium" wrote in message
...
You are referring to ~100 points difference. From what I've seen, that is
the margin of error/variaton, for those progs. +/- 100 points, in either,
is negligable.


The margin may be small, but the fact is they are still slower, and the
results are repeatable. I've seen a post here where somebody said he gained
600 points by changing to the 3.10's. I'm merely making the point that *did
not* occur here.

-
KCB stood up at show-n-tell, in ezuEb.396698$Dw6.1247555@attbi_s02, and
said:

3DMark2001 SE ---- 3.9 = 14580 3.10 = 14440

3DMark03 ---------- 3.9 = 4653 3.10 = 4582

9700 Pro - default settings for both drivers
1024*768*32 @ 85 MHz

MOH:AA had some weird texture problems in one map, but I think it was
caused by the mod that was running. Tried 4 other maps and all were
OK. COD worked perfectly, no glitches. Refresh override still works
good.

--
Strontium

"It's no surprise, to me. I am my own worst enemy. `Cause every
now, and then, I kick the livin' **** `outta me." - Lit

-
KCB stood up at show-n-tell, in y5AEb.427323$275.1303279@attbi_s53, and
said:

"Strontium" wrote in message
...
You are referring to ~100 points difference. From what I've seen,
that is the margin of error/variaton, for those progs. +/- 100
points, in either, is negligable.


The margin may be small, but the fact is they are still slower, and
the results are repeatable. I've seen a post here where somebody
said he gained 600 points by changing to the 3.10's. I'm merely
making the point that *did not* occur here.

And, I'm making the point that 100 points, whether repeatable or not, is not
worth spit. The benchmarks, you are using, are biased to begin with.
That's besides the point. A 100 point difference, plus or minus, is within
the tolerance level. If you want to judge the drivers, do it with your real
world use...unless, of course, that's all you do with your machine is run
biased, sponsored benchmarks...then have at it. 100 points, even then, is
still negligable.


snip
--
Strontium

"It's no surprise, to me. I am my own worst enemy. `Cause every
now, and then, I kick the livin' **** `outta me." - Lit

"KCB" wrote in message
news:y5AEb.427323$275.1303279@attbi_s53...

The margin may be small, but the fact is they are still slower, and the
results are repeatable. I've seen a post here where somebody said he
gained
600 points by changing to the 3.10's. I'm merely making the point that
*did
not* occur here.


You're talking about a synthetic benchmark here, and a DX8 one as well. Run
some *real-world* benchmarks with *real* games and see how fast the drivers
are. 3dmark is made for people who are too lazy or ignorant of how to
benchmark their system properly.

K

I wish the driver writers would quit worring about FPS and benchmarks and
keep focused on image quality, compatability and stability. What good are
a
few extra FPS if the system crashes or displays distorted textures? I
guess
it's the benchmarks and FPS that are cited in reviews and that sells
cards.

Yup. Nvidia set off that speed craziness when they were still fighting 3dfx,
because their image quality was miserable in comparison. Gamers took the
bait, and it became a major selling factor up to today. Actually I like the
way Ati handles this. They do aim for good IQ, but carefully enough not to
lose in the speed race. Objectively seen, everything above 60 FPS is a
waste. Unless the game is programmed sloppily, higher FPS become
unnoticeable. Any current card can achieve 60 FPS in any current game. And
also any current card has to drop below eventually in newer games. Yet even
30 FPS is still playable. Cards that are faster have a longer life, because
they drop below 60/30 later. Yet that isn't really in the interest of video
card companies, since they want to sell their newer models. But as long as
speed plays the major role it does, cards will have more speed than
necessary. Of course the companies make up for that with the insane prices
of their top end cards, because they won't sell a new card to buyers of
these for some time. I really don't think a Radeon 9800 XT is that much more
expensive to produce than a 9600 XT as the price suggests. But the 9600 will
require upgrading about a year earlier, so they can sell it much cheaper. I
have used my good old 3dfx Voodoo 5 5500 for full 3 years. An upgrade simply
wasn't necessary earlier. Up to UT 2003 it could handle every game at
playable FPS. Only now did I have to upgrade, because the V5 simply cannot
do DX 8.1/9 or anything higher than OpenGL 1.1. But still, 3 years of
service isn't bad at all, considering that there have been next to no driver
updates in all that time, except for a few fan-made ones.

What are you talking about? 3D Mark (all of them) are excellent games. I
play them all the time. ;-)

Brad

Word.

Stability and looks is IMHO more interesting than some xtra FPS. 3.10 works
great here. No probs at all.

9800PRO
P-4 2,6
1GB RAM
2X 120GB 7200 WD, 8MB
17" TFT
SB AUDIGY

/ C.

"Kent_Diego" skrev i meddelandet
news:yTuEb.37066$hf1.10764@lakeread06...
I wish the driver writers would quit worring about FPS and benchmarks and
keep focused on image quality, compatability and stability. What good are
a
few extra FPS if the system crashes or displays distorted textures? I
guess
it's the benchmarks and FPS that are cited in reviews and that sells
cards.

-Kent