9800 pro vs. 9700 pro

Fish

Distinguished
Apr 4, 2004
163
0
18,680
Archived from groups: alt.comp.periphs.videocards.ati (More info?)

I recently had my 9700pro damaged due to a poorly designed NB waterblock.
Long story short, a Swiftech northbridge block mounted on an Abit IC7 ripped
the holding cleats out of the board. I found it laying on top of my Radeon
9700pro along with all the mounting hardware. No water leaks but the video
card was having all sorts of problems and the system was running overclocked
without a NB cooler at all. Could have lost the motherboard too.

I replaced the 9700p with a Sapphire 9800pro from Newegg.com for $198. I
tried a ATI built 'refurbished' one for $160 but it was DOA.
I have had the 9700pro for about 2 full years and was planning on buying a
X800 of types in a few months when the price drops some more. But I was
forced into a replacement card now and the 9800pro is the best bargain.

Out of the box the 9800pro is getting BIG overclock numbers. The cards gets
really hot and I am planning on putting some good cooling on it myself.

What I noticed immediately is an improved 2D, richer and smoother colors. 3D
was even a bigger surprise, sharper 3D images, richer and more vibrant
colors. Very impressed.

I also noticed that for some games and for 3Dmark 2001 the performance jump
from the 9700pro and even the performance gains from overclocking the 9800p
was very small.
However, for some other games like UT2004 and with 3Dmark2003 the
performance jump was very significant! Also with Eye Candy like AA, the
9800pro leaps above the 9700pro.

9800pro with Cat 4.7 drivers, WinXP pro, P4 at 3260Mhz (250Mhz FSB):
3Dmark 2001: (default 1024x768x100Hz refresh)
18700 -- 378/338 (stock clock)
19400 -- 432/358 (No artifacts overclock on stock cooling)
** 18050 -- 9700p at 371/331 (No artifacts overclock)
- 14% overclock yielded a 4% performance gain.

3Dmark 2003: (default 1024x768x100Hz refresh)
5950 -- 378/338 (stock clock)
6550 -- 432/358 (No artifacts overclock on stock cooling)
6700 -- 452/364 (Shows artifacts but does not crash)
** 5400 -- 9700p at 371/331 (Shows artifacts!)
- 14% overclock yielded a 10% performance gain.

Aquamark3: (default 1024x768x100Hz refresh)
44,800 -- 378/338 (stock clock)
48,700 -- 432/358 (No artifacts overclock on stock cooling)
** 43,600 -- 9700p at 371/331 (No artifacts overclock)
- 14% overclock yielded a 9% performance gain.
Unreal Tournament 2004:
UT04 retail version, DM_Asbestos.ut2 bots, 12 bots

1280 X 1024 X 85Hz refresh.
100 fps (21min/215max) -- 378/338 -- No Eye Candy
101 fps (35min/218max) -- 432/358 -- No Eye Candy
61 fps (15min/132max) -- 378/338 -- 2x AA, 4x AF
71 fps (19min/134max) -- 432/358 -- 2x AA, 4x AF
41 fps (13min/115max) -- 378/338 -- 4x AA, 8x AF
43 fps (15min/120max) -- 432/358 -- 4x AA, 8x AF

1600 X 1200 X 85Hz refresh.
91 fps (21min/176max) -- 378/338 -- No Eye Candy
93 fps (31min/203max) -- 432/358 -- No Eye Candy
37 fps (20min/100max) -- 378/338 -- 2x AA, 4x AF
40 fps (20min/132max) -- 432/358 -- 2x AA, 4x AF

Conclusion for Unreal Tournament 2004 is that it shows consistency with
3Dmark 2001 for older game engines. Performance gains are minimal.
The only real advantage I can see, and I wish I had some numbers to show it,
is that at 1600 X 1200 UT04 was a bit to slow to play with my 9700pro. At
1280 X 1024 and a little Eye Candy game play was smooth and playable but not
a fast without any AA or AF.
The 9800pro is playing 1600 X 1200 with 2x AA and 4x AF just fine and quite
fast. That alone is a marked improvement along with the increased visual
quality making the 9800pro the deal of the year right now.

BTW: The 9800XT is just a factory overclocked 9800pro, isn't it?
 

Fish

Distinguished
Apr 4, 2004
163
0
18,680
Archived from groups: alt.comp.periphs.videocards.ati (More info?)

correction:

However, for some other games like UT2004 and with 3Dmark2003 the
> performance jump was very significant! Also with Eye Candy like AA, the
> 9800pro leaps above the 9700pro.

"fish" <fish@email.com> wrote in message
news:kEvMc.8391$YK2.2127700@news4.srv.hcvlny.cv.net...
> I recently had my 9700pro damaged due to a poorly designed NB waterblock.
> Long story short, a Swiftech northbridge block mounted on an Abit IC7
ripped
> the holding cleats out of the board. I found it laying on top of my Radeon
> 9700pro along with all the mounting hardware. No water leaks but the video
> card was having all sorts of problems and the system was running
overclocked
> without a NB cooler at all. Could have lost the motherboard too.
>
> I replaced the 9700p with a Sapphire 9800pro from Newegg.com for $198. I
> tried a ATI built 'refurbished' one for $160 but it was DOA.
> I have had the 9700pro for about 2 full years and was planning on buying a
> X800 of types in a few months when the price drops some more. But I was
> forced into a replacement card now and the 9800pro is the best bargain.
>
> Out of the box the 9800pro is getting BIG overclock numbers. The cards
gets
> really hot and I am planning on putting some good cooling on it myself.
>
> What I noticed immediately is an improved 2D, richer and smoother colors.
3D
> was even a bigger surprise, sharper 3D images, richer and more vibrant
> colors. Very impressed.
>
> I also noticed that for some games and for 3Dmark 2001 the performance
jump
> from the 9700pro and even the performance gains from overclocking the
9800p
> was very small.
> However, for some other games like UT2004 and with 3Dmark2003 the
> performance jump was very significant! Also with Eye Candy like AA, the
> 9800pro leaps above the 9700pro.
>
> 9800pro with Cat 4.7 drivers, WinXP pro, P4 at 3260Mhz (250Mhz FSB):
> 3Dmark 2001: (default 1024x768x100Hz refresh)
> 18700 -- 378/338 (stock clock)
> 19400 -- 432/358 (No artifacts overclock on stock cooling)
> ** 18050 -- 9700p at 371/331 (No artifacts overclock)
> - 14% overclock yielded a 4% performance gain.
>
> 3Dmark 2003: (default 1024x768x100Hz refresh)
> 5950 -- 378/338 (stock clock)
> 6550 -- 432/358 (No artifacts overclock on stock cooling)
> 6700 -- 452/364 (Shows artifacts but does not crash)
> ** 5400 -- 9700p at 371/331 (Shows artifacts!)
> - 14% overclock yielded a 10% performance gain.
>
> Aquamark3: (default 1024x768x100Hz refresh)
> 44,800 -- 378/338 (stock clock)
> 48,700 -- 432/358 (No artifacts overclock on stock cooling)
> ** 43,600 -- 9700p at 371/331 (No artifacts overclock)
> - 14% overclock yielded a 9% performance gain.
> Unreal Tournament 2004:
> UT04 retail version, DM_Asbestos.ut2 bots, 12 bots
>
> 1280 X 1024 X 85Hz refresh.
> 100 fps (21min/215max) -- 378/338 -- No Eye Candy
> 101 fps (35min/218max) -- 432/358 -- No Eye Candy
> 61 fps (15min/132max) -- 378/338 -- 2x AA, 4x AF
> 71 fps (19min/134max) -- 432/358 -- 2x AA, 4x AF
> 41 fps (13min/115max) -- 378/338 -- 4x AA, 8x AF
> 43 fps (15min/120max) -- 432/358 -- 4x AA, 8x AF
>
> 1600 X 1200 X 85Hz refresh.
> 91 fps (21min/176max) -- 378/338 -- No Eye Candy
> 93 fps (31min/203max) -- 432/358 -- No Eye Candy
> 37 fps (20min/100max) -- 378/338 -- 2x AA, 4x AF
> 40 fps (20min/132max) -- 432/358 -- 2x AA, 4x AF
>
> Conclusion for Unreal Tournament 2004 is that it shows consistency with
> 3Dmark 2001 for older game engines. Performance gains are minimal.
> The only real advantage I can see, and I wish I had some numbers to show
it,
> is that at 1600 X 1200 UT04 was a bit to slow to play with my 9700pro. At
> 1280 X 1024 and a little Eye Candy game play was smooth and playable but
not
> a fast without any AA or AF.
> The 9800pro is playing 1600 X 1200 with 2x AA and 4x AF just fine and
quite
> fast. That alone is a marked improvement along with the increased visual
> quality making the 9800pro the deal of the year right now.
>
> BTW: The 9800XT is just a factory overclocked 9800pro, isn't it?
>
>
>
 

Fish

Distinguished
Apr 4, 2004
163
0
18,680
Archived from groups: alt.comp.periphs.videocards.ati (More info?)

correction:

"However, for some other games like UT2004 and with 3Dmark2003 the
performance jump was very significant!"

I meant to type; "...for some other games like DOOM3 and with 3Dmark2003 the
performance jump was very significant!"

Sorry for the confusion.


"fish" <fish@email.com> wrote in message
news:kEvMc.8391$YK2.2127700@news4.srv.hcvlny.cv.net...
> I recently had my 9700pro damaged due to a poorly designed NB waterblock.
> Long story short, a Swiftech northbridge block mounted on an Abit IC7
ripped
> the holding cleats out of the board. I found it laying on top of my Radeon
> 9700pro along with all the mounting hardware. No water leaks but the video
> card was having all sorts of problems and the system was running
overclocked
> without a NB cooler at all. Could have lost the motherboard too.
>
> I replaced the 9700p with a Sapphire 9800pro from Newegg.com for $198. I
> tried a ATI built 'refurbished' one for $160 but it was DOA.
> I have had the 9700pro for about 2 full years and was planning on buying a
> X800 of types in a few months when the price drops some more. But I was
> forced into a replacement card now and the 9800pro is the best bargain.
>
> Out of the box the 9800pro is getting BIG overclock numbers. The cards
gets
> really hot and I am planning on putting some good cooling on it myself.
>
> What I noticed immediately is an improved 2D, richer and smoother colors.
3D
> was even a bigger surprise, sharper 3D images, richer and more vibrant
> colors. Very impressed.
>
> I also noticed that for some games and for 3Dmark 2001 the performance
jump
> from the 9700pro and even the performance gains from overclocking the
9800p
> was very small.
> However, for some other games like UT2004 and with 3Dmark2003 the
> performance jump was very significant! Also with Eye Candy like AA, the
> 9800pro leaps above the 9700pro.
>
> 9800pro with Cat 4.7 drivers, WinXP pro, P4 at 3260Mhz (250Mhz FSB):
> 3Dmark 2001: (default 1024x768x100Hz refresh)
> 18700 -- 378/338 (stock clock)
> 19400 -- 432/358 (No artifacts overclock on stock cooling)
> ** 18050 -- 9700p at 371/331 (No artifacts overclock)
> - 14% overclock yielded a 4% performance gain.
>
> 3Dmark 2003: (default 1024x768x100Hz refresh)
> 5950 -- 378/338 (stock clock)
> 6550 -- 432/358 (No artifacts overclock on stock cooling)
> 6700 -- 452/364 (Shows artifacts but does not crash)
> ** 5400 -- 9700p at 371/331 (Shows artifacts!)
> - 14% overclock yielded a 10% performance gain.
>
> Aquamark3: (default 1024x768x100Hz refresh)
> 44,800 -- 378/338 (stock clock)
> 48,700 -- 432/358 (No artifacts overclock on stock cooling)
> ** 43,600 -- 9700p at 371/331 (No artifacts overclock)
> - 14% overclock yielded a 9% performance gain.
> Unreal Tournament 2004:
> UT04 retail version, DM_Asbestos.ut2 bots, 12 bots
>
> 1280 X 1024 X 85Hz refresh.
> 100 fps (21min/215max) -- 378/338 -- No Eye Candy
> 101 fps (35min/218max) -- 432/358 -- No Eye Candy
> 61 fps (15min/132max) -- 378/338 -- 2x AA, 4x AF
> 71 fps (19min/134max) -- 432/358 -- 2x AA, 4x AF
> 41 fps (13min/115max) -- 378/338 -- 4x AA, 8x AF
> 43 fps (15min/120max) -- 432/358 -- 4x AA, 8x AF
>
> 1600 X 1200 X 85Hz refresh.
> 91 fps (21min/176max) -- 378/338 -- No Eye Candy
> 93 fps (31min/203max) -- 432/358 -- No Eye Candy
> 37 fps (20min/100max) -- 378/338 -- 2x AA, 4x AF
> 40 fps (20min/132max) -- 432/358 -- 2x AA, 4x AF
>
> Conclusion for Unreal Tournament 2004 is that it shows consistency with
> 3Dmark 2001 for older game engines. Performance gains are minimal.
> The only real advantage I can see, and I wish I had some numbers to show
it,
> is that at 1600 X 1200 UT04 was a bit to slow to play with my 9700pro. At
> 1280 X 1024 and a little Eye Candy game play was smooth and playable but
not
> a fast without any AA or AF.
> The 9800pro is playing 1600 X 1200 with 2x AA and 4x AF just fine and
quite
> fast. That alone is a marked improvement along with the increased visual
> quality making the 9800pro the deal of the year right now.
>
> BTW: The 9800XT is just a factory overclocked 9800pro, isn't it?
>
>
>
 

Fish

Distinguished
Apr 4, 2004
163
0
18,680
Archived from groups: alt.comp.periphs.videocards.ati (More info?)

This is interesting: Radeon 9700p at 371/344 - 3320 (255fsb), DirectX 8,
Cat3.4

19,300 score (http://service.futuremark.com/compare?2k1=6661809)





Ever since DX9 and the CAT 4.x drivers, DX8 games have gotten a big bit
taken out of them.







"fish" <fish@email.com> wrote in message
news:kEvMc.8391$YK2.2127700@news4.srv.hcvlny.cv.net...
> I recently had my 9700pro damaged due to a poorly designed NB waterblock.
> Long story short, a Swiftech northbridge block mounted on an Abit IC7
ripped
> the holding cleats out of the board. I found it laying on top of my Radeon
> 9700pro along with all the mounting hardware. No water leaks but the video
> card was having all sorts of problems and the system was running
overclocked
> without a NB cooler at all. Could have lost the motherboard too.
>
> I replaced the 9700p with a Sapphire 9800pro from Newegg.com for $198. I
> tried a ATI built 'refurbished' one for $160 but it was DOA.
> I have had the 9700pro for about 2 full years and was planning on buying a
> X800 of types in a few months when the price drops some more. But I was
> forced into a replacement card now and the 9800pro is the best bargain.
>
> Out of the box the 9800pro is getting BIG overclock numbers. The cards
gets
> really hot and I am planning on putting some good cooling on it myself.
>
> What I noticed immediately is an improved 2D, richer and smoother colors.
3D
> was even a bigger surprise, sharper 3D images, richer and more vibrant
> colors. Very impressed.
>
> I also noticed that for some games and for 3Dmark 2001 the performance
jump
> from the 9700pro and even the performance gains from overclocking the
9800p
> was very small.
> However, for some other games like UT2004 and with 3Dmark2003 the
> performance jump was very significant! Also with Eye Candy like AA, the
> 9800pro leaps above the 9700pro.
>
> 9800pro with Cat 4.7 drivers, WinXP pro, P4 at 3260Mhz (250Mhz FSB):
> 3Dmark 2001: (default 1024x768x100Hz refresh)
> 18700 -- 378/338 (stock clock)
> 19400 -- 432/358 (No artifacts overclock on stock cooling)
> ** 18050 -- 9700p at 371/331 (No artifacts overclock)
> - 14% overclock yielded a 4% performance gain.
>
> 3Dmark 2003: (default 1024x768x100Hz refresh)
> 5950 -- 378/338 (stock clock)
> 6550 -- 432/358 (No artifacts overclock on stock cooling)
> 6700 -- 452/364 (Shows artifacts but does not crash)
> ** 5400 -- 9700p at 371/331 (Shows artifacts!)
> - 14% overclock yielded a 10% performance gain.
>
> Aquamark3: (default 1024x768x100Hz refresh)
> 44,800 -- 378/338 (stock clock)
> 48,700 -- 432/358 (No artifacts overclock on stock cooling)
> ** 43,600 -- 9700p at 371/331 (No artifacts overclock)
> - 14% overclock yielded a 9% performance gain.
> Unreal Tournament 2004:
> UT04 retail version, DM_Asbestos.ut2 bots, 12 bots
>
> 1280 X 1024 X 85Hz refresh.
> 100 fps (21min/215max) -- 378/338 -- No Eye Candy
> 101 fps (35min/218max) -- 432/358 -- No Eye Candy
> 61 fps (15min/132max) -- 378/338 -- 2x AA, 4x AF
> 71 fps (19min/134max) -- 432/358 -- 2x AA, 4x AF
> 41 fps (13min/115max) -- 378/338 -- 4x AA, 8x AF
> 43 fps (15min/120max) -- 432/358 -- 4x AA, 8x AF
>
> 1600 X 1200 X 85Hz refresh.
> 91 fps (21min/176max) -- 378/338 -- No Eye Candy
> 93 fps (31min/203max) -- 432/358 -- No Eye Candy
> 37 fps (20min/100max) -- 378/338 -- 2x AA, 4x AF
> 40 fps (20min/132max) -- 432/358 -- 2x AA, 4x AF
>
> Conclusion for Unreal Tournament 2004 is that it shows consistency with
> 3Dmark 2001 for older game engines. Performance gains are minimal.
> The only real advantage I can see, and I wish I had some numbers to show
it,
> is that at 1600 X 1200 UT04 was a bit to slow to play with my 9700pro. At
> 1280 X 1024 and a little Eye Candy game play was smooth and playable but
not
> a fast without any AA or AF.
> The 9800pro is playing 1600 X 1200 with 2x AA and 4x AF just fine and
quite
> fast. That alone is a marked improvement along with the increased visual
> quality making the 9800pro the deal of the year right now.
>
> BTW: The 9800XT is just a factory overclocked 9800pro, isn't it?
>
>
>
 
G

Guest

Guest
Archived from groups: alt.comp.periphs.videocards.ati (More info?)

fish wrote:
> BTW: The 9800XT is just a factory overclocked 9800pro, isn't it?

The 9800 Pro is R350 core.
The XT is R360 core.

The R360 has a temperature sensor.

Otherwise pretty much the same, I think.

Performance-wise, the XT is an overclocked pro.

Ben
--
A7N8X FAQ: www.ben.pope.name/a7n8x_faq.html
Questions by email will likely be ignored, please use the newsgroups.
I'm not just a number. To many, I'm known as a String...
 
G

Guest

Guest
Archived from groups: alt.comp.periphs.videocards.ati (More info?)

In article <kEvMc.8391$YK2.2127700@news4.srv.hcvlny.cv.net>,
fish@email.com (fish) wrote:

> But I was forced into a replacement card now and the 9800pro
> is the best bargain.

FWIW subtract the cost of the 9800Pro from the cost of an X800 and ask
yourself if you'd pay *that* price for an X800 if that's what it cost now.
Because in real terms that's what it *is* costing you... though at least
this way you get a spare 9800Pro to sell on.

I've often made the mistake of spending less to save money, but when I
look back I realise I could have had better hardware, sooner for the same
overall expenditure.

Andrew McP
 

Fish

Distinguished
Apr 4, 2004
163
0
18,680
Archived from groups: alt.comp.periphs.videocards.ati (More info?)

The problem is that the X800Pro isn't the card you want. Its the X800XTPE
that you want and that's still close to $500.

X800 Pro -- (R360), 110million transistors, 412MHz clock, 900MHz memory 4
vertex pipelines, 8 pixel pipelines, 2.0 shaders, .15 fab.

X800 XP PE -- (R420), 160million transistors, 500MHz clock, 1120MHz memory 6
vertex pipelines, 16 pixel pipelines, 2.0+ shders, .13 fab.


Sure, I can sell the 9800Pro for at least $100 by the end of the year or
give it to my kids. My Kids PC is an XP2400+, 9600Pro and my wife's PC is an
XP2200+ with a 9600se. Either will benefit from the 9800pro.



"Andrew MacPherson" <andrew.mcp@DELETETHISdsl.pipex.com> wrote in message
news:memo.20040724213651.2792C@address_disguised.address_disguised...
> In article <kEvMc.8391$YK2.2127700@news4.srv.hcvlny.cv.net>,
> fish@email.com (fish) wrote:
>
> > But I was forced into a replacement card now and the 9800pro
> > is the best bargain.
>
> FWIW subtract the cost of the 9800Pro from the cost of an X800 and ask
> yourself if you'd pay *that* price for an X800 if that's what it cost now.
> Because in real terms that's what it *is* costing you... though at least
> this way you get a spare 9800Pro to sell on.
>
> I've often made the mistake of spending less to save money, but when I
> look back I realise I could have had better hardware, sooner for the same
> overall expenditure.
>
> Andrew McP
>
 
G

Guest

Guest
Archived from groups: alt.comp.periphs.videocards.ati (More info?)

fish wrote:
> The problem is that the X800Pro isn't the card you want. Its the X800XTPE
> that you want and that's still close to $500.
>
> X800 Pro -- (R360), 110million transistors, 412MHz clock, 900MHz memory 4
> vertex pipelines, 8 pixel pipelines, 2.0 shaders, .15 fab.

Are you sure?

I think you'll the X800 Pro is an R420, 160M transistors, with 6 vertex and
12 pixel pipelines and 0.13u process running at 475MHz with 900MHz GDDR3
memory.

Sounds like you described the 9800XT.

Ben
--
A7N8X FAQ: www.ben.pope.name/a7n8x_faq.html
Questions by email will likely be ignored, please use the newsgroups.
I'm not just a number. To many, I'm known as a String...
 

Fish

Distinguished
Apr 4, 2004
163
0
18,680
Archived from groups: alt.comp.periphs.videocards.ati (More info?)

http://www.anandtech.com/video/showdoc.aspx?i=2044&p=2

? Could be wrong.



"Ben Pope" <spam@hotmail.com> wrote in message
news:2mg78qFmbr6kU1@uni-berlin.de...
> fish wrote:
> > The problem is that the X800Pro isn't the card you want. Its the
X800XTPE
> > that you want and that's still close to $500.
> >
> > X800 Pro -- (R360), 110million transistors, 412MHz clock, 900MHz memory
4
> > vertex pipelines, 8 pixel pipelines, 2.0 shaders, .15 fab.
>
> Are you sure?
>
> I think you'll the X800 Pro is an R420, 160M transistors, with 6 vertex
and
> 12 pixel pipelines and 0.13u process running at 475MHz with 900MHz GDDR3
> memory.
>
> Sounds like you described the 9800XT.
>
> Ben
> --
> A7N8X FAQ: www.ben.pope.name/a7n8x_faq.html
> Questions by email will likely be ignored, please use the newsgroups.
> I'm not just a number. To many, I'm known as a String...
>
>
 

Fish

Distinguished
Apr 4, 2004
163
0
18,680
Archived from groups: alt.comp.periphs.videocards.ati (More info?)

Yes, old data on Anandtech.

http://www.bjorn3d.com/read.php?cID=634

You are correct.

I don't like that the X800Pro only has 12 pipelines.
I still won't buy a Pro for $350+ , I'll wait and get a XTPE when its about
$300.
OR an NV40 which is beating the snot out of ATI's best in 'newer' tech games
like Doom3. ID has already announced a slew of games coming out that will be
built on the Carmak Doom3 engine. Apparently the pixilshaders on the NV40
are superior to the R420.

I have had a Voodoo3, Voodoo3500, GF2, GF3, GF4 Ti4600's as well as an 7000,
8500, 9700pro and now this 9800pro.
I guess that makes me crazy but its fun.


"fish" <fish@email.com> wrote in message
news:4MBMc.9589$YK2.3289569@news4.srv.hcvlny.cv.net...
> http://www.anandtech.com/video/showdoc.aspx?i=2044&p=2
>
> ? Could be wrong.
>
>
>
> "Ben Pope" <spam@hotmail.com> wrote in message
> news:2mg78qFmbr6kU1@uni-berlin.de...
> > fish wrote:
> > > The problem is that the X800Pro isn't the card you want. Its the
> X800XTPE
> > > that you want and that's still close to $500.
> > >
> > > X800 Pro -- (R360), 110million transistors, 412MHz clock, 900MHz
memory
> 4
> > > vertex pipelines, 8 pixel pipelines, 2.0 shaders, .15 fab.
> >
> > Are you sure?
> >
> > I think you'll the X800 Pro is an R420, 160M transistors, with 6 vertex
> and
> > 12 pixel pipelines and 0.13u process running at 475MHz with 900MHz GDDR3
> > memory.
> >
> > Sounds like you described the 9800XT.
> >
> > Ben
> > --
> > A7N8X FAQ: www.ben.pope.name/a7n8x_faq.html
> > Questions by email will likely be ignored, please use the newsgroups.
> > I'm not just a number. To many, I'm known as a String...
> >
> >
>
>
 
G

Guest

Guest
Archived from groups: alt.comp.periphs.videocards.ati (More info?)

fish wrote:
> Yes, old data on Anandtech.

No, two separate tables that you failed to separate.

> http://www.bjorn3d.com/read.php?cID=634
>
> You are correct.
>
> I don't like that the X800Pro only has 12 pipelines.

Only? I bet thats more than your current GPU.

> I still won't buy a Pro for $350+ , I'll wait and get a XTPE when its
> about $300.

And then it'll have "only" 16 pipelines, 'cos the R500 (or whatever) will
have 24 pipelines! Draw the line somewhere.

> OR an NV40 which is beating the snot out of ATI's best in 'newer' tech
> games like Doom3.

You of course mean, "different tech" (openGL) vs DirectX, where the ATI
performs rather well. The "other" game: HL2.

Don't forget that the ATI and nVidia philosophy are different. Doom3 was
developed on nVidia (ok so the nVidia has a better architecture/support for
stencil shadows), and that HL2 was developed on ATI. That makes a HUGE
difference! Just take a look at Anandtechs recent delve into console
emulation - Pete develops on ATI and hence ATI wipes the floor with nVidia.
If they are going to start factoring that into their benchmarks they'd best
be careful to explain that detail at the bottom of EVERY page with the
scores or my opinion of Anandtech will drop immensely. My point is that due
to the different architectures, you can easily bias yourself one way or the
other with a few simple decisions at the beginning of development.

> ID has already announced a slew of games coming out
> that will be built on the Carmak Doom3 engine. Apparently the
> pixilshaders on the NV40 are superior to the R420.

Maybe, haven't really compared the two in enough detail to comment - I do
know they perform similarly in the benchmarks I've seen - it's a lot closer
than the previous generation and I'm glad that nVidia have working Pixel
Shaders now. But R420 is a stop gap. It's a trial migration to 0.13micron
and doesn't have the features they originally planned for R420.

If you hadn't noticed this is a game of hurdles, nVidia are in front, ATI
are in front, nVidia are in front, ATI are in front... if it wasn't like
that one of the companies would fold. In recent times ATI have had the edge
for a higher proportion of the time. I have no loyalty either way, I've
never tried an nVidia I went from a V3 3000 on a PIII 450 to my 9800 Pro on
a Barton 2500+ and have been satisfied. When the time comes to upgrade I
will do my research and choose the card thats best for me at the time.

Same for the Processor. Hurdles. Intel have been pulling some silly tricks
to get ahead - such as the EE which they can't even make for a sensible
price and is hot enough to heat a small house.

This is barely about technology any more, it's bloody marketing.

> I have had a Voodoo3, Voodoo3500, GF2, GF3, GF4 Ti4600's as well as an
> 7000, 8500, 9700pro and now this 9800pro.
> I guess that makes me crazy but its fun.

If you hadn't mentioned all them ATI cards I'd have thought you were
trolling with your previous comments.

Hmm, time to do some work, and stop talking technology I think!

Ben
--
A7N8X FAQ: www.ben.pope.name/a7n8x_faq.html
Questions by email will likely be ignored, please use the newsgroups.
I'm not just a number. To many, I'm known as a String...
 

Fish

Distinguished
Apr 4, 2004
163
0
18,680
Archived from groups: alt.comp.periphs.videocards.ati (More info?)

The Anandtec article doesn't mention the 9800XP. Re-read the article, the
information is outdated!

Anandtech said: '...vertex engine of R420 includes 6 total vertex pipelines
(R3xx has four). This gives R420 a 50% per clock increase in peak vertex
shader power per clock cycle...'
Old news thats incorrect today.

The games I play (and they are few since I am too busy to play them all)
will primarily be on the Doom3 engine platform (Quake4, Wolfenstein, Doom3)
so I may actually benefit from an NV40. Thats all I'm saying.
ID games tend to be the 'most popular' games and the ones I like to play. If
they run faster with nVidia then I'll buy nVidia, simple as that.


"Ben Pope" <spam@hotmail.com> wrote in message
news:2mgb0gFmv4paU1@uni-berlin.de...
> fish wrote:
> > Yes, old data on Anandtech.
>
> No, two separate tables that you failed to separate.
>
> > http://www.bjorn3d.com/read.php?cID=634
> >
> > You are correct.
> >
> > I don't like that the X800Pro only has 12 pipelines.
>
> Only? I bet thats more than your current GPU.
>
> > I still won't buy a Pro for $350+ , I'll wait and get a XTPE when its
> > about $300.
>
> And then it'll have "only" 16 pipelines, 'cos the R500 (or whatever) will
> have 24 pipelines! Draw the line somewhere.
>
> > OR an NV40 which is beating the snot out of ATI's best in 'newer' tech
> > games like Doom3.
>
> You of course mean, "different tech" (openGL) vs DirectX, where the ATI
> performs rather well. The "other" game: HL2.
>
> Don't forget that the ATI and nVidia philosophy are different. Doom3 was
> developed on nVidia (ok so the nVidia has a better architecture/support
for
> stencil shadows), and that HL2 was developed on ATI. That makes a HUGE
> difference! Just take a look at Anandtechs recent delve into console
> emulation - Pete develops on ATI and hence ATI wipes the floor with
nVidia.
> If they are going to start factoring that into their benchmarks they'd
best
> be careful to explain that detail at the bottom of EVERY page with the
> scores or my opinion of Anandtech will drop immensely. My point is that
due
> to the different architectures, you can easily bias yourself one way or
the
> other with a few simple decisions at the beginning of development.
>
> > ID has already announced a slew of games coming out
> > that will be built on the Carmak Doom3 engine. Apparently the
> > pixilshaders on the NV40 are superior to the R420.
>
> Maybe, haven't really compared the two in enough detail to comment - I do
> know they perform similarly in the benchmarks I've seen - it's a lot
closer
> than the previous generation and I'm glad that nVidia have working Pixel
> Shaders now. But R420 is a stop gap. It's a trial migration to
0.13micron
> and doesn't have the features they originally planned for R420.
>
> If you hadn't noticed this is a game of hurdles, nVidia are in front, ATI
> are in front, nVidia are in front, ATI are in front... if it wasn't like
> that one of the companies would fold. In recent times ATI have had the
edge
> for a higher proportion of the time. I have no loyalty either way, I've
> never tried an nVidia I went from a V3 3000 on a PIII 450 to my 9800 Pro
on
> a Barton 2500+ and have been satisfied. When the time comes to upgrade I
> will do my research and choose the card thats best for me at the time.
>
> Same for the Processor. Hurdles. Intel have been pulling some silly
tricks
> to get ahead - such as the EE which they can't even make for a sensible
> price and is hot enough to heat a small house.
>
> This is barely about technology any more, it's bloody marketing.
>
> > I have had a Voodoo3, Voodoo3500, GF2, GF3, GF4 Ti4600's as well as an
> > 7000, 8500, 9700pro and now this 9800pro.
> > I guess that makes me crazy but its fun.
>
> If you hadn't mentioned all them ATI cards I'd have thought you were
> trolling with your previous comments.
>
> Hmm, time to do some work, and stop talking technology I think!
>
> Ben
> --
> A7N8X FAQ: www.ben.pope.name/a7n8x_faq.html
> Questions by email will likely be ignored, please use the newsgroups.
> I'm not just a number. To many, I'm known as a String...
>
>
 
G

Guest

Guest
Archived from groups: alt.comp.periphs.videocards.ati (More info?)

I had two Woodoo 2 12mb in SLI, and before that, the original Woodoo of
course...
Yes its crazy and still great fun!
Btw, the price on the Woodoo2 was around 340 Euros. Each ;-o

Martin.

>
> I have had a Voodoo3, Voodoo3500, GF2, GF3, GF4 Ti4600's as well as an
7000,
> 8500, 9700pro and now this 9800pro.
> I guess that makes me crazy but its fun.
>
 
G

Guest

Guest
Archived from groups: alt.comp.periphs.videocards.ati (More info?)

In article <iWBMc.9612$YK2.3319300@news4.srv.hcvlny.cv.net>,
fish@email.com (fish) wrote:

> I don't like that the X800Pro only has 12 pipelines.

At least up until now if you buy the VIVO version of the Pro you're
getting a "slightly failed" XT VIVO. With a little messing about you can
have the extra pipelines, though not quite XT speeds.

Of course it's always possible that the disabled pipelines have flaws
which led to the "Pro-isation" of the card, but so far it usually seems to
have been down to failures to hit XT core speeds reliably. I certainly see
no artefacts in my XT-ised Pro, and it now benchmarks in XT territory. The
difference isn;t remarkably, but it's worth having for free.

As ever though, you can't rely on this stuff, and I guess there *is* a
cost because invalidating the warranty on new kit isn't very clever :)

Andrew "not very clever" McP
 
G

Guest

Guest
Archived from groups: alt.comp.periphs.videocards.ati (More info?)

fish wrote:
> The Anandtec article doesn't mention the 9800XP. Re-read the article, the
> information is outdated!
>
> Anandtech said: '...vertex engine of R420 includes 6 total vertex
> pipelines (R3xx has four). This gives R420 a 50% per clock increase in
> peak vertex shader power per clock cycle...'
> Old news thats incorrect today.

You're a muppet.

12 pixel pipelines. 6 vertex pipelines. Like I said in the first place.

It's correct. Now go away.

> The games I play (and they are few since I am too busy to play them all)
> will primarily be on the Doom3 engine platform (Quake4, Wolfenstein,
> Doom3) so I may actually benefit from an NV40. Thats all I'm saying.
> ID games tend to be the 'most popular' games and the ones I like to play.
> If they run faster with nVidia then I'll buy nVidia, simple as that.

Fine.

Ben
--
A7N8X FAQ: www.ben.pope.name/a7n8x_faq.html
Questions by email will likely be ignored, please use the newsgroups.
I'm not just a number. To many, I'm known as a String...