Performance differences with GF4 MX440 cards

I'm an occasional system builder and for a while now have been fitting MSI
(Micro Star Int.) GF4 64Mb DDR AGP 8X cards to systems whose users require
basic gaming capabilities. Flight simulator, etc. I've been sourcing these
for about £35 ($50) and they are fanless, easy to install, stable and
reliable.

I've just picked up a similar Abit GF4 MX-8X E with 64Mb DDR for only £19
($30) and tried a quick 3Dmark2001SE benchmark between the two systems. In
fact the Abit card was in a better specified PC; Athlon 2500+, 512MB PC2700
and 8X AGP MSI KT4AV motherboard, whilst the MSI card was in a 1.GHz Duron,
256MB PC2100 and a AGP 4X Syntax motherboard.

Both systems were running Windows XP (with latest patches) with motherboard
chipset drivers loaded and the latest version of nVidia drivers.

MSI get 4,658 3Dmarks and the Abit Siluro gets 3,112. Why the large
difference?

The cards look the same, both small format without fans. Yet when I look on
benchmark web-sites and see that some GF4MX cards can do 7,500 3D marks!

I see from this news group that there are SE version of this nVidia chipset,
but I can't see what are the differences. Perhaps this is the performance
factor.

Help me choose a better 3D graphics card to stock . I see that the FX5200
is available for £50 ($80) can I justify the extra cost to my customers?
Any recommendations on the manufacturer and model?

Jonathan

The MX card 440 will not perform all the test in 3dmark2001, that is why
even thoouh you might see the same performance, it will get a lower score.
i have no idea what card I would tell you to stock but I would say that
I have a Fx5200 card for now and it is a good starter card for todays games.

"Jonathan Eales" wrote in message
...
I'm an occasional system builder and for a while now have been fitting MSI
(Micro Star Int.) GF4 64Mb DDR AGP 8X cards to systems whose users require
basic gaming capabilities. Flight simulator, etc. I've been sourcing
these
for about £35 ($50) and they are fanless, easy to install, stable and
reliable.

I've just picked up a similar Abit GF4 MX-8X E with 64Mb DDR for only £19
($30) and tried a quick 3Dmark2001SE benchmark between the two systems.
In
fact the Abit card was in a better specified PC; Athlon 2500+, 512MB
PC2700
and 8X AGP MSI KT4AV motherboard, whilst the MSI card was in a 1.GHz
Duron,
256MB PC2100 and a AGP 4X Syntax motherboard.

Both systems were running Windows XP (with latest patches) with
motherboard
chipset drivers loaded and the latest version of nVidia drivers.

MSI get 4,658 3Dmarks and the Abit Siluro gets 3,112. Why the large
difference?

The cards look the same, both small format without fans. Yet when I look
on
benchmark web-sites and see that some GF4MX cards can do 7,500 3D marks!

I see from this news group that there are SE version of this nVidia
chipset,
but I can't see what are the differences. Perhaps this is the performance
factor.

Help me choose a better 3D graphics card to stock . I see that the FX5200
is available for £50 ($80) can I justify the extra cost to my customers?
Any recommendations on the manufacturer and model?

Jonathan

Jonathan Eales wrote:
I'm an occasional system builder and for a while now have been
fitting MSI (Micro Star Int.) GF4 64Mb DDR AGP 8X cards to systems
whose users require basic gaming capabilities. Flight simulator,
etc. I've been sourcing these for about £35 ($50) and they are
fanless, easy to install, stable and reliable.

I've just picked up a similar Abit GF4 MX-8X E with 64Mb DDR for only
£19 ($30) and tried a quick 3Dmark2001SE benchmark between the two
systems. In fact the Abit card was in a better specified PC; Athlon
2500+, 512MB PC2700 and 8X AGP MSI KT4AV motherboard, whilst the MSI
card was in a 1.GHz Duron, 256MB PC2100 and a AGP 4X Syntax
motherboard.

Both systems were running Windows XP (with latest patches) with
motherboard chipset drivers loaded and the latest version of nVidia
drivers.

MSI get 4,658 3Dmarks and the Abit Siluro gets 3,112. Why the large
difference?

The cards look the same, both small format without fans. Yet when I
look on benchmark web-sites and see that some GF4MX cards can do
7,500 3D marks!

I see from this news group that there are SE version of this nVidia
chipset, but I can't see what are the differences. Perhaps this is
the performance factor.

Help me choose a better 3D graphics card to stock . I see that the
FX5200 is available for £50 ($80) can I justify the extra cost to my
customers? Any recommendations on the manufacturer and model?

FX5200 should be ok for basic gaming but its basically the same as a GF4
MX440 with DX9.0.

If I were building it, I would go for a 9600 Pro or a Ti4200 but then again,
it's just for basic gaming so a FX5200 should be ok.

But why do supposedly similar model GF4 MX440 cards from different
manufacturers perform so differently? And do I switch to MSI's FX5200
because their GF4 MX440 is faster?

Thanks,
Jonathan


"MongolHunter" wrote in message
...
Jonathan Eales wrote:
I'm an occasional system builder and for a while now have been
fitting MSI (Micro Star Int.) GF4 64Mb DDR AGP 8X cards to systems
whose users require basic gaming capabilities. Flight simulator,
etc. I've been sourcing these for about £35 ($50) and they are
fanless, easy to install, stable and reliable.

I've just picked up a similar Abit GF4 MX-8X E with 64Mb DDR for only
£19 ($30) and tried a quick 3Dmark2001SE benchmark between the two
systems. In fact the Abit card was in a better specified PC; Athlon
2500+, 512MB PC2700 and 8X AGP MSI KT4AV motherboard, whilst the MSI
card was in a 1.GHz Duron, 256MB PC2100 and a AGP 4X Syntax
motherboard.

Both systems were running Windows XP (with latest patches) with
motherboard chipset drivers loaded and the latest version of nVidia
drivers.

MSI get 4,658 3Dmarks and the Abit Siluro gets 3,112. Why the large
difference?

The cards look the same, both small format without fans. Yet when I
look on benchmark web-sites and see that some GF4MX cards can do
7,500 3D marks!

I see from this news group that there are SE version of this nVidia
chipset, but I can't see what are the differences. Perhaps this is
the performance factor.

Help me choose a better 3D graphics card to stock . I see that the
FX5200 is available for £50 ($80) can I justify the extra cost to my
customers? Any recommendations on the manufacturer and model?

FX5200 should be ok for basic gaming but its basically the same as a GF4
MX440 with DX9.0.

If I were building it, I would go for a 9600 Pro or a Ti4200 but then
again,
it's just for basic gaming so a FX5200 should be ok.

"Jonathan Eales" wrote in message
...
I'm an occasional system builder and for a while now have been fitting MSI
(Micro Star Int.) GF4 64Mb DDR AGP 8X cards to systems whose users require
basic gaming capabilities. Flight simulator, etc. I've been sourcing
these
for about £35 ($50) and they are fanless, easy to install, stable and
reliable.

I've just picked up a similar Abit GF4 MX-8X E with 64Mb DDR for only £19
($30) and tried a quick 3Dmark2001SE benchmark between the two systems.
In
fact the Abit card was in a better specified PC; Athlon 2500+, 512MB
PC2700
and 8X AGP MSI KT4AV motherboard, whilst the MSI card was in a 1.GHz
Duron,
256MB PC2100 and a AGP 4X Syntax motherboard.

Both systems were running Windows XP (with latest patches) with
motherboard
chipset drivers loaded and the latest version of nVidia drivers.

MSI get 4,658 3Dmarks and the Abit Siluro gets 3,112. Why the large
difference?

Before jumping to conclusions try swapping the cards to see if it is
anything to do with driver settings - if the difference is still evident
then I would suspect lower core/mem speeds in the slower card and/or reduced
spec such as 64 bit rather than 128 bit memory (what does that 'E' at the
end of the Abit GF4 MX-8X E indicate????).

SNIP

Help me choose a better 3D graphics card to stock . I see that the FX5200
is available for £50 ($80) can I justify the extra cost to my customers?
Any recommendations on the manufacturer and model?

An ATI Radeon 9000 will provide similar performance to an FX5200 and might
also be worth considering. Ebuyer currently have the 64 MB Sapphire 9000
(fanless) for £44.64 inc. However, it is becoming increasingly difficult to
compare different graphics cards because of the plethora of different core
speeds and memory speeds which, together with variations in the amount of
onboard memory (64/128/256) and the variations in data pathways
(64bit/128bit/256bit), make it very hard to predict performance with
confidence unless the detailed spec of the particular card is available.
Make sure you avoid anything that is 'SE' or 'lite' or has anything else in
its name or description that indicates a spec below the standard for the
chipset.

Tony

Tony, thanks for this.

I have done a bit of extra experimentation. Yes, the MSI card did run even
slightly more faster (AGP 8X effect?) when the cards were swapped over so
the driver settings were consistent.

I then discovered the nVidia 'coolbits' registry setting allowing me to
examine and adjust the core and memory speeds of the cards and the Abit card
was significantly lower in both. When adjusted to the equivalent setting of
the MSI card the benchmark ran just slightly slower on the Abit card. So
your suspicions of lower core/memory speeds were correct. Trying to push
both cards faster quickly caused weird coloured dots on the screen. Case
temperatures also rose quickly. So I reduced to more optimum values.

I researched your recommendation of the ATI Radeon 9000 card only to find
that is DirectX 8 compared to the DirectX 9 support offered by the FX5200.
However, whether this will be noticeable or important to part-time gamers I
cannot say. I will order one of these cards from ebuyer as well as a MSI
Fx5200 8x AGP 128MB which costs £52 inc. and then test the two against the
cheaper GF4 MX440 cards.

I'm happy that I discovered the reason for the differences. Neither card is
a MX440SE version as both work with AGP 8X and the SE apparently doesn't
support that.

Thanks again for your help.

Jonathan

"TMack" wrote in message
. ..

"Jonathan Eales" wrote in message
...
I'm an occasional system builder and for a while now have been fitting
MSI
(Micro Star Int.) GF4 64Mb DDR AGP 8X cards to systems whose users
require
basic gaming capabilities. Flight simulator, etc. I've been sourcing
these
for about £35 ($50) and they are fanless, easy to install, stable and
reliable.

I've just picked up a similar Abit GF4 MX-8X E with 64Mb DDR for only
£19
($30) and tried a quick 3Dmark2001SE benchmark between the two systems.
In
fact the Abit card was in a better specified PC; Athlon 2500+, 512MB
PC2700
and 8X AGP MSI KT4AV motherboard, whilst the MSI card was in a 1.GHz
Duron,
256MB PC2100 and a AGP 4X Syntax motherboard.

Both systems were running Windows XP (with latest patches) with
motherboard
chipset drivers loaded and the latest version of nVidia drivers.

MSI get 4,658 3Dmarks and the Abit Siluro gets 3,112. Why the large
difference?

Before jumping to conclusions try swapping the cards to see if it is
anything to do with driver settings - if the difference is still evident
then I would suspect lower core/mem speeds in the slower card and/or
reduced
spec such as 64 bit rather than 128 bit memory (what does that 'E' at the
end of the Abit GF4 MX-8X E indicate????).

SNIP

Help me choose a better 3D graphics card to stock . I see that the
FX5200
is available for £50 ($80) can I justify the extra cost to my customers?
Any recommendations on the manufacturer and model?

An ATI Radeon 9000 will provide similar performance to an FX5200 and might
also be worth considering. Ebuyer currently have the 64 MB Sapphire 9000
(fanless) for £44.64 inc. However, it is becoming increasingly difficult
to
compare different graphics cards because of the plethora of different core
speeds and memory speeds which, together with variations in the amount of
onboard memory (64/128/256) and the variations in data pathways
(64bit/128bit/256bit), make it very hard to predict performance with
confidence unless the detailed spec of the particular card is available.
Make sure you avoid anything that is 'SE' or 'lite' or has anything else
in
its name or description that indicates a spec below the standard for the
chipset.

Tony

"Jonathan Eales" wrote in message
...
I researched your recommendation of the ATI Radeon 9000 card only to find
that is DirectX 8 compared to the DirectX 9 support offered by the
FX5200.

Arguably, DX9 support is wasted on the FX5200, as it's too slow anyway.

Alex

On Thu, 20 Nov 2003 10:08:45 -0000, "Jonathan Eales"
wrote:

Tony, thanks for this.

I have done a bit of extra experimentation. Yes, the MSI card did run even
slightly more faster (AGP 8X effect?) when the cards were swapped over so
the driver settings were consistent.

I then discovered the nVidia 'coolbits' registry setting allowing me to
examine and adjust the core and memory speeds of the cards and the Abit card
was significantly lower in both. When adjusted to the equivalent setting of
the MSI card the benchmark ran just slightly slower on the Abit card. So
your suspicions of lower core/memory speeds were correct. Trying to push
both cards faster quickly caused weird coloured dots on the screen. Case
temperatures also rose quickly. So I reduced to more optimum values.

I researched your recommendation of the ATI Radeon 9000 card only to find
that is DirectX 8 compared to the DirectX 9 support offered by the FX5200.
However, whether this will be noticeable or important to part-time gamers I
cannot say. I will order one of these cards from ebuyer as well as a MSI
Fx5200 8x AGP 128MB which costs £52 inc. and then test the two against the
cheaper GF4 MX440 cards.

I'm happy that I discovered the reason for the differences. Neither card is
a MX440SE version as both work with AGP 8X and the SE apparently doesn't
support that.

MX440se are AGP 8x cards... they are actually 420 cores with improved
memory.... but sitll crap.


--
Remember when real men used Real computers!?
When 512K of video RAM was a lot!

Death to Palladium & WPA!!