• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

About how much better is a 9600 GSO than a x1950 pro?

Emanon Retsim

New Member
Joined
Jun 30, 2009
Messages
13 (0.00/day)
Heyas folks. I'm looking to get a few more miles outta of my system before I do a full rebuild hopefully next spring. Right now my old x1950 is on the fritz and needs to be replaced with something cheap and decent. Newegg is selling Asus 9600 GSO 384MB for $35 bucks after rebate. Should I snag it while it's on sell or go with something else?

http://www.newegg.com/Product/Product.aspx?Item=N82E16814121319


Right now my system for gaming at 1280x1024 is;
Mobo: MSI K8N Neo4-F Socket 939
CPU: Single Core AMD Athlon 64 3800+
Ram: DDR1 2x1gb G.Skill
GPU: Sapphire Radeon x1950 Pro
OS: Vista Home Premium 32bit
 
I noticed about double the frames in games comparing my x1950xtx (48 shaders) OC'd vs my Palit 9600GT stock in many games I played, from CoD4, BF2142, Supreme Commander, etc. The GSO is better on shader power with 96 vs my GT's 64, though the GT's shaders are faster. I miss my old Palit, that card clocked like a demon too. But you should see a very nice increase, though 384MB isn't too much for gaming at 1280x1024 you should be fine for the most part, though you might want to consider spending a tad more if youc an find a 768MB version. But it shouldn't be all too necessary. For that price it's definately worth it imo, especially over the aged 1950pro, it'll be a night and day difference, with decent DX10 gaming at that resolution to boot.

:toast:
 
you wont see much of a increase because ur cpu is bottlenecking the rig but at least a decent increase u will se.
 
I agree with the above posters.i used a 7900GS OC(about same as x1950pro) and i went to a 8800GT i saw a big jump in frame rates and the benchmark scores are more then 2x but i wouldn't really bother getting a new card yet i would get a new CPU first.... BUT at that good of a deal i would snag it
 
you wont see much of a increase because ur cpu is bottlenecking the rig but at least a decent increase u will se.

I agree with the above posters.i used a 7900GS OC(about same as x1950pro) and i went to a 8800GT i saw a big jump in frame rates and the benchmark scores are more then 2x but i wouldn't really bother getting a new card yet i would get a new CPU first.... BUT at that good of a deal i would snag it

If he can already max it out, yeah, the cpu will be a bottleneck, but at that point, what does it matter? He already has good frames.

Now, on games that his card struggles, he'll notice quite a large improvement in frames. With most modern games, the gpu is the bottleneck, not the cpu. The cpu won't make any difference if the card is being pushed past it's limits.

He should absolutely buy the card first, and worry about a cpu later. DaMulta proved this by throwing an 8800GT into a single core Athlon system back when we were still sponsored.
 
can u get the street fighter 4 benchmark and run it and tell us what it scores, Its a good bench
 
Heyas folks. I'm looking to get a few more miles outta of my system before I do a full rebuild hopefully next spring. Right now my old x1950 is on the fritz and needs to be replaced with something cheap and decent. Newegg is selling Asus 9600 GSO 384MB for $35 bucks after rebate. Should I snag it while it's on sell or go with something else?

http://www.newegg.com/Product/Product.aspx?Item=N82E16814121319


Right now my system for gaming at 1280x1024 is;
Mobo: MSI K8N Neo4-F Socket 939
CPU: Single Core AMD Athlon 64 3800+
Ram: DDR1 2x1gb G.Skill
GPU: Sapphire Radeon x1950 Pro
OS: Vista Home Premium 32bit

Hey Emanon
I am in the same situation . My Asus x1950 pro has become "Heat Weak" , so
I just purchased the above card from Newegg just before i found this thread.
http://www.newegg.com/Product/Product.aspx?Item=N82E16814121320
Someone on another forum suggested upgrading the fan even before installing it first time.

Please Let me know how the card works out for you and your experience with it.
Thanks
Dave

Mobo: Asus A8n32-SLI Deluxe, socket 939
CPU: Athlon 64 4000+
RAM: 1 GB Kingston HYPERX DDR434
GPU: ASUS x1950 pro 256 DDR3
PSU: Antec TruePower Trio 650w
OS: XPpro sp3
 
Ok here are the results.
it was a bit of a let down.:ohwell:
Just looking at the card , it was warped and the heat sink was on a bit cockeyed. Solder joints looked a bit messy. Visably, it doesnt seem to have the same quality as the 385mb version, although, I understand Asus gets their board from different various manufactures.
O well it works.;)

Although the 9600GSO does allow me to finish a game at 1600x1200, gameplay is much choppier than that of the x1950 pro,, go figure :confused:
I suspect it may have something to do with how the RAM is being used.

Here at the 3Dmark05 scores at 1600x1200
x1950pro 7518
9600GSO 9100

Oddly enough ,, With the 9600GSO the 3Dmark simulations were much choppier than the x1950pro, yet it had a higher frame rate,, can't figure that out :wtf:
The new Geforce card seems to be putting off just as much heat as the ATI yet isn't shutting down.
So i guess it was a bit of a trade off,, It works ;)
 
those 9600GSO's overclock like crazy.....
 
the 9600 is a worthwhile upgrade over the 1950(especially if the 1950 is the 256mb version) It will work fine with your 3800. (no noticeable bottleneck, especially at your res)

I just built a rig with a pair of those GSO's in SLI and it kicks ass. If possible look for the 512mb version.

I have used MANY of the 9600s and really love them, they are the sweet spot for value (best perf,lowest power/heat for the $$)

although I have a pair of 9800GTs I wish I had kept my last pair of 9600GTs.
 
Last edited:
Ok here are the results.
it was a bit of a let down.:ohwell:
Just looking at the card , it was warped and the heat sink was on a bit cockeyed. Solder joints looked a bit messy. Visably, it doesnt seem to have the same quality as the 385mb version, although, I understand Asus gets their board from different various manufactures.
O well it works.;)

Although the 9600GSO does allow me to finish a game at 1600x1200, gameplay is much choppier than that of the x1950 pro,, go figure :confused:
I suspect it may have something to do with how the RAM is being used.

Here at the 3Dmark05 scores at 1600x1200
x1950pro 7518
9600GSO 9100

Oddly enough ,, With the 9600GSO the 3Dmark simulations were much choppier than the x1950pro, yet it had a higher frame rate,, can't figure that out :wtf:
The new Geforce card seems to be putting off just as much heat as the ATI yet isn't shutting down.
So i guess it was a bit of a trade off,, It works ;)
Don't forget to try different driver versions. Some perform better than others.
 
btw..i have a BFG Tech 9400GT 512megs DDR2 Pci-e 16X 2.0
what is the tool that i can use to overclock it ?
i know i could get more from that card. thx
 
have you heard the expression " you cant get blood out of a rock":laugh:

RivaTuner is what I use for my cards
 
you should have looked at what you were buying, that 9600gso only has 8 rops meaning the pixel fill rate is alot lower than a standard gso and is only 128bit meaning it has way less memory bandwidth

this is your gso on gpu z
GPUZ.png

this is a stock gso
gpuz.gif


just to confirm, does your card look like this one
DSCN0285.jpg


and are you sure this is the one you bought
http://www.newegg.com/Product/Product.aspx?Item=N82E16814121320

those are matts pics by the way
 
Last edited:
Found out that this 9600GSO only has a 128bit memory bus. There are MANY flavors of 9600GSO. With 128/192/256bit memory bus, 384/512/768/1GB ddr2 OR DDR3 RAM, different ROP count, and shader counts from 96 to 48, there have been many cards calling themselves 9600GSO.
 
Found out that this 9600GSO only has a 128bit memory bus. There are MANY flavors of 9600GSO. With 128/192/256bit memory bus, 384/512/768/1GB ddr2 OR DDR3 RAM, different ROP count, and shader counts from 96 to 48, there have been many cards calling themselves 9600GSO.
Thats the problem when nVidia remarked the 8800GS to 9600GSo and later slap on a G94 and refer it by the same name :laugh:

its so fucked up man, newegg is selling 8rop, 128 bit 9600gso's now and one of them is a 48sp model meaning its barely faster than a 9500GT, but all have 8 rops and are 128bit now.
http://www.newegg.com/Product/Produ...48 106792522 1067939752&name=GeForce 9600 GSO
thats just so messed up

wth happened

48 sp, 8rop gso sucks
Well so you actually have no idea that the guy is buying the "standard" GSO?
Next, time when you recommand somebody to buy a nV card, do some research on nVidia's naming mess :nutkick:
 
Last edited:
None the less, nVidia's low end right now are full of ripoffs, I wouldn't recommand anyone buying those POS.

That G92 crap is one of those later GSOs that are castrated to match the G94 fail fests.:respect:

The G94 GSO actually performs much better than this G92. :shadedshu
 
Last edited:
dave bought that card before he was even on this thread, I think the thread op got the good one though,. . . . . I hope

Sorry Dave,

sorry
 
He brought that junk off newegg, by any chance he can return it and get his money back?
 
Oddly enough ,, With the 9600GSO the 3Dmark simulations were much choppier than the x1950pro, yet it had a higher frame rate,, can't figure that out :wtf:
The new Geforce card seems to be putting off just as much heat as the ATI yet isn't shutting down.
So i guess it was a bit of a trade off,, It works ;)

Keep in mind that a 3Dmark test is a torture test. The program is designed to push your card to the edge to see how much it will take to make the card crap out. As such, cards will vary in their response to be pushed to the edge. Some start getting really chopping, some ghost, some start freezing up, etc. For a benchmark test, just go with the final score and not really how it looked during the test.

This difference, I think can also be caused by new options like PhysX being enabled for the new card that were not available at first, because the X1950 can't do those function.
 
Keep in mind that a 3Dmark test is a torture test. The program is designed to push your card to the edge to see how much it will take to make the card crap out. As such, cards will vary in their response to be pushed to the edge. Some start getting really chopping, some ghost, some start freezing up, etc. For a benchmark test, just go with the final score and not really how it looked during the test.

This difference, I think can also be caused by new options like PhysX being enabled for the new card that were not available at first, because the X1950 can't do those function.
The "offical" 3DMark Vantage score should have GPU Physx disabled.
Which really puts the load on his CPU.

Vantage don't even run on X1950s.
So I will assume that he is running 3Dmark06:
In DirectX 9.0c those ROPs on the X1950 really comes in handy, and those 96 unified shaders on that GSO really doesn't do much shit.
Now that I look it up, the X1950Pro have almost double memory bandwidth, 12 vs 8 ROPs and even a higher pixel fillrate... :banghead:

The X1950Pro actually held its ground against the 8600GTS, so don't know what to say about this 9600GSO...:shadedshu
I guess you got DX10. :respect:
 
Last edited:
ummm, no

dave posted a 3dmark 05 score of about 9K, and even then he says it was outperformed his ripped down gso, now I have a gs which is the same as a original gso and it scores 15K stock, and those extra sp's do mean something and you can't compare pixel fill rate or texture rate to any card Geforce 8/Radeon HD2000 or higher, because the architecture is not the same.

I had a X1950XTX and I can tell you right now, this little crappy 8800GS greatly outperformed it when OC'd

the G94 GSO is not faster then the G92 GSO, I had both to replace a GS that died on me just to find out nvidia got me, the card is slower than a normal 96sp gso.
 
Last edited:
Back
Top