is the only difference the clock and memory speed?
like if you o/c a Ti-200 to 240/500 clock/memory, do u basically have a Ti-500?
like if you o/c a Ti-200 to 240/500 clock/memory, do u basically have a Ti-500?
Difference between GF3 Ti-200 and Ti-500?
- Thread starter pookguy88
- Start date
The Ti 200 has the Geforce3 cores that couldn't pass the muster at 240 stable. Now it might be true that Nvidia only puts cores in the Ti 500 that are stable at 250 (to give some leeway), and the ones that aren't are put in the Ti 200 bin. So you may get lucky. But most guys are getting maybe 220 max.
One of the reasons I went all out and got a Ti 500 is because no matter how good an overclock you do on the Ti 200 to equal or exceed a Ti 500, the Ti 500 can overclock too.
Case in point: I just bought an Asus 8200-T5 Deluxe card a couple of days ago. It's a Ti 500, and just for fun, I decided to see how high it will go on stock cooling. How about 266/603 stable? And when I mean stable, I mean it gets zero 2D artifacts at that setting, no mean feat. I use Artifact Tester 5.0's software test, and also the proprietary 2D test that came with my Asus card to check this. I can clock even higher (270/601) and be stable in 3D, but it won't pass these very sensitive 2D artifact tests.
One of the reasons I went all out and got a Ti 500 is because no matter how good an overclock you do on the Ti 200 to equal or exceed a Ti 500, the Ti 500 can overclock too.
Case in point: I just bought an Asus 8200-T5 Deluxe card a couple of days ago. It's a Ti 500, and just for fun, I decided to see how high it will go on stock cooling. How about 266/603 stable? And when I mean stable, I mean it gets zero 2D artifacts at that setting, no mean feat. I use Artifact Tester 5.0's software test, and also the proprietary 2D test that came with my Asus card to check this. I can clock even higher (270/601) and be stable in 3D, but it won't pass these very sensitive 2D artifact tests.
hmm ok, well I clocked my Ti-200 to 220/500 and it was completely stable in 3dMark2001 for like 3 hours but it won't bonk out on me in Q3A somtimes, anyone know whats up with that?
where might I get this artifact tester??
thanx.
where might I get this artifact tester??
thanx.
You can get it here:
http://www.voodoofiles.com/type.asp
Let me tell you, you are gonna have a hard time getting past the Software test with 0 artifacts. It is very sensitive.
When I say I got my card stable at 266/603 I did this:
1. Ran 3DMark2000 and checked very closely for faint snow in the Helicopter sequence on HIGH settings. Then checked for tiny flashing textures in the Advernture sequence on HIGH settings. You have to look very closely, and if you blink, you'll miss it. But it should be absolutely clean. Another was is to check QuakeIII in a game, and stare at a really complex scene. Look for the tiny bits of black flash.
2. Ran 3DMark2001 and made sure it didn't hang on the transitions.
3. Ran a real demanding QuakeIII demo, like The Reverend's Quaver demo, all settings maxed. Ran it twice, at 1280x1024 and 1024x768. You don't want to lock up here, or see the screen go bonkers with black triangles.
4. Ran Artifact Tester 5.0's DirectX 7 test on HARDCORE settings. Got 0 artifacts.
5. Ran Artifact Tester 5.0's Software test. Got 0 artifacts. You might pass the DirectX test but not get by this one with 0 artifacts. At a 606 memory clock I get maybe 47 artifacts, and at 609 I get 217.
If you really want to be a sadist with your video card, you can try running the Vulpine demo and see if it will complete it OK.
Some guys don't believe I broke 600 on the memory, but the tests I ran speak for itself. I did this on stock cooling and room temp of 21 C. The temp of the card stayed around 53 C. This Asus card I got kick's ass!
http://www.voodoofiles.com/type.asp
Let me tell you, you are gonna have a hard time getting past the Software test with 0 artifacts. It is very sensitive.
When I say I got my card stable at 266/603 I did this:
1. Ran 3DMark2000 and checked very closely for faint snow in the Helicopter sequence on HIGH settings. Then checked for tiny flashing textures in the Advernture sequence on HIGH settings. You have to look very closely, and if you blink, you'll miss it. But it should be absolutely clean. Another was is to check QuakeIII in a game, and stare at a really complex scene. Look for the tiny bits of black flash.
2. Ran 3DMark2001 and made sure it didn't hang on the transitions.
3. Ran a real demanding QuakeIII demo, like The Reverend's Quaver demo, all settings maxed. Ran it twice, at 1280x1024 and 1024x768. You don't want to lock up here, or see the screen go bonkers with black triangles.
4. Ran Artifact Tester 5.0's DirectX 7 test on HARDCORE settings. Got 0 artifacts.
5. Ran Artifact Tester 5.0's Software test. Got 0 artifacts. You might pass the DirectX test but not get by this one with 0 artifacts. At a 606 memory clock I get maybe 47 artifacts, and at 609 I get 217.
If you really want to be a sadist with your video card, you can try running the Vulpine demo and see if it will complete it OK.
Some guys don't believe I broke 600 on the memory, but the tests I ran speak for itself. I did this on stock cooling and room temp of 21 C. The temp of the card stayed around 53 C. This Asus card I got kick's ass!
Clevor,
I got 0 artifacts for both software and directx artifact tester at 220/500, it says "zero artifacts, o/c some more"
I got 0 artifacts for both software and directx artifact tester at 220/500, it says "zero artifacts, o/c some more"
"How do you get your puter to run the Quake III demo in a loop?"
Uhhh, I don't loop a $365 video card at 603 memory clock hours on end! Some guys believe in running 3DMark2001 overnight to test stability, but you are getting more 3D wear-and tear out of that card than in 3 years of normal use. Though I don't know if wear-and-tear apply to video cards like a car engine.
There are ways of looping a QuakeIII demo but I've never done it. You can loop 3DMark2001 pretty easy if you want to. But the Quaver demo is about as demanding as it gets to test overclocks.
And from Pookguy88's experience, looks like you can get zero 2D artifacts but still lock up in 3D.
Last night I decided I HAD TO KNOW. One last test. So I ran the Vulpine demo at 1280x1024 with 64-tap anisotrophy and Quincunx on, and the card passed at 266/603. I only got 27 fps, but then I am still running an 800 Thunderbird and 100 FSB.
I won't leave the card at 266/603 until I put on aftermarket sinks and fan, or at least a video card cooler with 2-80MM fans blowing over the card. I'm thinking an Alpha copper insert heatsink (low profile) I got in a junk bin here in Japan for $9 for the CPU. This Alpha heatsink is for the Socket A and lists for $43! For the ram, Alpha forged aluminum heatsinks thermal epoxied to 1-1.5 mm copper shims. I have a Voodoo5 in the PCI slot that is running 44 C max under stress with a copper insert heatsink and fan. The Asus is running around 53-58 C with the stock cooling.
Uhhh, I don't loop a $365 video card at 603 memory clock hours on end! Some guys believe in running 3DMark2001 overnight to test stability, but you are getting more 3D wear-and tear out of that card than in 3 years of normal use. Though I don't know if wear-and-tear apply to video cards like a car engine.
There are ways of looping a QuakeIII demo but I've never done it. You can loop 3DMark2001 pretty easy if you want to. But the Quaver demo is about as demanding as it gets to test overclocks.
And from Pookguy88's experience, looks like you can get zero 2D artifacts but still lock up in 3D.
Last night I decided I HAD TO KNOW. One last test. So I ran the Vulpine demo at 1280x1024 with 64-tap anisotrophy and Quincunx on, and the card passed at 266/603. I only got 27 fps, but then I am still running an 800 Thunderbird and 100 FSB.
I won't leave the card at 266/603 until I put on aftermarket sinks and fan, or at least a video card cooler with 2-80MM fans blowing over the card. I'm thinking an Alpha copper insert heatsink (low profile) I got in a junk bin here in Japan for $9 for the CPU. This Alpha heatsink is for the Socket A and lists for $43! For the ram, Alpha forged aluminum heatsinks thermal epoxied to 1-1.5 mm copper shims. I have a Voodoo5 in the PCI slot that is running 44 C max under stress with a copper insert heatsink and fan. The Asus is running around 53-58 C with the stock cooling.
That's a nice OC on the Ti500 clevor. I'm currently running my Visiontek GF3 Ti500 at 260/590. I have run it at 265/600 before, but it seems like when I oc my fsb it won't run stably at that speed.
well, i found my limit i think. I got it to 200/500 with 0 artifacts (software and directx) and can run everything stablely now...i guess its not that bad of a jump from stock 175/400 speed.
Clevor the video cards like expectancy is like 10 years of continuous use under normal operating conditions, I highly doubt that runing 3dmark overnight will put 3 yesrs of wear and tare on it....
In case ya'll haven't read the article on the TI's at Firing Squad, here's a quote:
"At $200, the GeForce3 Ti 200 certainly looks attractive to us. Sure, it isn't quite able to keep up with the Ti 500, but remember that it costs 43% less. Unless you're able to find a GeForce3 board for a similar price, we stand behind our belief that the GeForce3 Ti 200 is the way to go for the consumer that wants a DirectX 8 board but is on a limited budget. Once the Ti 200 is overclocked, it becomes quite a performer. "
IMHO if you're worried at all about your budget then the TI 200 is the way to go, but if either you're a die hard gamer or have googles of money and want the best then look at the TI 500. Hope that helps.
Scoob
"At $200, the GeForce3 Ti 200 certainly looks attractive to us. Sure, it isn't quite able to keep up with the Ti 500, but remember that it costs 43% less. Unless you're able to find a GeForce3 board for a similar price, we stand behind our belief that the GeForce3 Ti 200 is the way to go for the consumer that wants a DirectX 8 board but is on a limited budget. Once the Ti 200 is overclocked, it becomes quite a performer. "
IMHO if you're worried at all about your budget then the TI 200 is the way to go, but if either you're a die hard gamer or have googles of money and want the best then look at the TI 500. Hope that helps.
Scoob
A geforce3 ti200 clocked the same as a ti500 WILL NOT perform the same. The ti500 has a memory bandwidth of 8.0GB/Sec. A ti200 has a memory bandwidth of 6.4GB/Sec. This way Nvidia isn't selling the exact same cards (ti200, GF3, ti500) with nothing different than clock speeds. The memory bandwidth is different so you will buy the ti500. It's kinda like have some sdram and ddr ram running at the same FSB....
These are some really intresting tests you guys have done... whats the secret to O/cing a vid card? Where do i go to change the mhz settings etc?
I have another thread i started somewhere asking what software i need to o/c etc, if anyone can tell me about this ill be a true happy camper
Card i got is a PNY geforce3 ti 200, for the normal price of $199.. at first i thought i got jipped cuz i didnt see any diffrence in games like wolfenstien, but then i checked my settings and it was all turned off to low settings.. once i got it all set to high etc, WOW i was totally amazed.
Somehow id still like to know how to O/C this thing if its possible
can anyone help me?
I have another thread i started somewhere asking what software i need to o/c etc, if anyone can tell me about this ill be a true happy camper
Card i got is a PNY geforce3 ti 200, for the normal price of $199.. at first i thought i got jipped cuz i didnt see any diffrence in games like wolfenstien, but then i checked my settings and it was all turned off to low settings.. once i got it all set to high etc, WOW i was totally amazed.
Somehow id still like to know how to O/C this thing if its possible
can anyone help me?
<< A geforce3 ti200 clocked the same as a ti500 WILL NOT perform the same. The ti500 has a memory bandwidth of 8.0GB/Sec. A ti200 has a memory bandwidth of 6.4GB/Sec. >>
Aren't the memory bandwidth specs based on the clock speed of the memory? Up the clock = more data flows through per second = higher bandwidth?
In fact, looking at the numbers... if you take the stock memory clock of each card, (500/400), and then compare it to the bandwidth difference (8.0/6.4) you'll see they are the exact same ratio
6.4/8 = 0.80
4/5 = 0.80
Makes sense, no?
<< A geforce3 ti200 clocked the same as a ti500 WILL NOT perform the same. The ti500 has a memory bandwidth of 8.0GB/Sec. A ti200 has a memory bandwidth of 6.4GB/Sec. This way Nvidia isn't selling the exact same cards (ti200, GF3, ti500) with nothing different than clock speeds. The memory bandwidth is different so you will buy the ti500. It's kinda like have some sdram and ddr ram running at the same FSB.... >>
They will perform the same at the same clockspeed. The Ti200's reported memory bandwidth of 6.4GB/s is because the ram is running at 400mhz. If you increase the ram speed on a Ti200 to 500mhz, then it will have 8.0GB/s of memory bandwidth like the Ti500. If you increase the gpu speed on the Ti200 to 240mhz and have the ram running at 500mhz then it will perform exactly the same as a Ti500 card running at stock 240/500 speed.
A good quality Ti200 can outperform a poor quality Ti500, and save you $200 besides.
When clocked identically, cards will perform identically.
Tested 3 "quality" Ti200's (GW GS, $175 delivered, NE) with stock cooling.
All hit at least 260 core and 560 memory.
(maxed at 262 core 561 memory, 3dmark 2001 error free).
Some low quality Ti500's will not hit 260/560.
When clocked identically, cards will perform identically.
Tested 3 "quality" Ti200's (GW GS, $175 delivered, NE) with stock cooling.
All hit at least 260 core and 560 memory.
(maxed at 262 core 561 memory, 3dmark 2001 error free).
Some low quality Ti500's will not hit 260/560.
i finally found out how to O/C my card and did a few tests.
Set the clock speed from 175/400 up to 220/500 in a few tries after testing 3dmark2001 and that artifact 5.0 on Hardcore.
All the tests pass fine and i get zero artifacts in hardcore test mode. I seem to be at the end line here, i cant go beyond 220/500 !
does this mean im O/Cing this card to the max? Theres nothing that will make me beable to go beyond this point?
Well i guess if this IS the limit of the card, atleast its running great.
btw, if your curious which card i have i looked it up.. its a PNY GF3 ti200 i got it for $199. i guess this was the standard price cuz i couldnt find it cheaper anywhere else
Set the clock speed from 175/400 up to 220/500 in a few tries after testing 3dmark2001 and that artifact 5.0 on Hardcore.
All the tests pass fine and i get zero artifacts in hardcore test mode. I seem to be at the end line here, i cant go beyond 220/500 !
does this mean im O/Cing this card to the max? Theres nothing that will make me beable to go beyond this point?
Well i guess if this IS the limit of the card, atleast its running great.
btw, if your curious which card i have i looked it up.. its a PNY GF3 ti200 i got it for $199. i guess this was the standard price cuz i couldnt find it cheaper anywhere else
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 21K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter