Infinite loop error?valkyre said:Apparently mine works fine with both the extra pipelines and vertex shaders unlocked, however I recently (after a month or two or running with everything unlocked just fine) encountered the "infinite loop" error, which caused me to reinstall due to me not being able to figure out how to fix it.
So I don't know if that had anything to do with it or not.
![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
Unlocking 6800 pipelines with RivaTuner
- Thread starter Kckazdude
- Start date
I´ve got an Albatron 6800nu, I can unlock both the pipelines and vertex unit, however the extra pipelines seem to be slightly damaged (artifacts only visible in the Dusk demo, not in 3Dmark03/05,Far Cry, Doom3). I disabled the softmod again, because I don´t really need it yet.
However, I am wondering if 350mhz are totally safe on the 6800nu core. My temperatures seem to be lower at 350mhz, than they are @ stock 325mhz, and all together it seems that I´d stress my card less when I´d oc it to 350mhz (both 2D and 3D):
2D:
325mhz @ idle = 53c
350mhz @ idle = 45c
3D:
325@maxload = 61c
350@maxload = 61c
By the way, I have a XP2400@2700, am I highly CPU bottlenecked and are these 3Dmark scores normal?
3dmark03 - 7847 (325/700 66.93@Q)
3dmark03 - 8280 (350/700 66.93@Q)
3dmark05 - 3171 (325/700 66.93@Q)
3dmark05 - 3327 (350/700 66.93@Q)
Thanks for helping out a noob!
However, I am wondering if 350mhz are totally safe on the 6800nu core. My temperatures seem to be lower at 350mhz, than they are @ stock 325mhz, and all together it seems that I´d stress my card less when I´d oc it to 350mhz (both 2D and 3D):
2D:
325mhz @ idle = 53c
350mhz @ idle = 45c
3D:
325@maxload = 61c
350@maxload = 61c
By the way, I have a XP2400@2700, am I highly CPU bottlenecked and are these 3Dmark scores normal?
3dmark03 - 7847 (325/700 66.93@Q)
3dmark03 - 8280 (350/700 66.93@Q)
3dmark05 - 3171 (325/700 66.93@Q)
3dmark05 - 3327 (350/700 66.93@Q)
Thanks for helping out a noob!
karl76 said:I´ve got an Albatron 6800nu, I can unlock both the pipelines and vertex unit, however the extra pipelines seem to be slightly damaged (artifacts only visible in the Dusk demo, not in 3Dmark03/05,Far Cry, Doom3). I disabled the softmod again, because I don´t really need it yet.
However, I am wondering if 350mhz are totally safe on the 6800nu core. My temperatures seem to be lower at 350mhz, than they are @ stock 325mhz, and all together it seems that I´d stress my card less when I´d oc it to 350mhz (both 2D and 3D):
2D:
325mhz @ idle = 53c
350mhz @ idle = 45c
3D:
325@maxload = 61c
350@maxload = 61c
By the way, I have a XP2400@2700, am I highly CPU bottlenecked and are these 3Dmark scores normal?
3dmark03 - 7847 (325/700 66.93@Q)
3dmark03 - 8280 (350/700 66.93@Q)
3dmark05 - 3171 (325/700 66.93@Q)
3dmark05 - 3327 (350/700 66.93@Q)
Thanks for helping out a noob!
Your scores are the same as mine, so it looks right to me.
Nazo
2[H]4U
- Joined
- Apr 2, 2002
- Messages
- 3,672
Ryland said:
I seriously doubt it, but, uhm, why do you ask rather than try? It's a simple matter to change the setting in the bios... Turning fastwrites on or off hasn't broken any systems that I know of.
I've had mine unlocked for a while now and have noticed a wierd problem.
Occasionally, in games, my system will hiccup. Out of nowhere the game will just pause and I can't do anything for maybe 5 seconds, then the game comes back. Everything that was happening continued, and I lose 5 seconds.
I only have the pipes unlocked, and this seems to happen whether I'm overclocked to 375/800 or at 325/700. I'm not sure, but I think this happened once or twice when I was at 12 pipes too. I don't get any artifacts and aside from this my system is rock solid. I have noticed that my temps seem a bit high, idling in the 60 to 70 range with an ambient of 35-45.
Using coolbits "detect optimal settings" feature, it said I should put my card at 389/807. Also, I don't know if this means anything, but I found it odd that when coolbits tested the settings it put my refresh back to 60Hz.
Could this hiccuping be caused by the unlock, or a card that runs too hot, or something else?
Occasionally, in games, my system will hiccup. Out of nowhere the game will just pause and I can't do anything for maybe 5 seconds, then the game comes back. Everything that was happening continued, and I lose 5 seconds.
I only have the pipes unlocked, and this seems to happen whether I'm overclocked to 375/800 or at 325/700. I'm not sure, but I think this happened once or twice when I was at 12 pipes too. I don't get any artifacts and aside from this my system is rock solid. I have noticed that my temps seem a bit high, idling in the 60 to 70 range with an ambient of 35-45.
Using coolbits "detect optimal settings" feature, it said I should put my card at 389/807. Also, I don't know if this means anything, but I found it odd that when coolbits tested the settings it put my refresh back to 60Hz.
Could this hiccuping be caused by the unlock, or a card that runs too hot, or something else?
Nazo
2[H]4U
- Joined
- Apr 2, 2002
- Messages
- 3,672
Normally the "hiccup" is textures/models being loaded into video memory. It's at it's absolute worst when the FSB and memory clocks aren't syncronized, agp should be at it's highest, etc. The idea is that you want to get all of the latencies involved in the process as low as you can. Unfortunately, it seems like in some games even the most powerful systems will stutter a tad. Most games are efficient and load as much as they can into the system memory, but, a few even end up keeping very little in real memory and just load from the harddrive (ugh.) Neverwinter Nights is an example of this. Even raising the max memory setting doesn't seem to make it use more, just less. (I usually catch it using a mere 32MB. Bear in mind, I have 1GB with always > 700MB free for games/etc.)
In your case, I'd say most especially look at the HTT. I guess higher is better for smoother gaming. If not that, I guess memory latencies are the next most important thing. Unfortunately, if it's the game itself, you're stuck.
In your case, I'd say most especially look at the HTT. I guess higher is better for smoother gaming. If not that, I guess memory latencies are the next most important thing. Unfortunately, if it's the game itself, you're stuck.
It happens in pretty much any game. I've had it happen in 3dmark05 and 03 as well. Only in games and benchmarks though. I'm going to do a little more testing, but it seemed like the higher my overclock the more frequently it happened, that's why I think it has something to do with the card itself. Perhaps I'm wrong though.
Nazo
2[H]4U
- Joined
- Apr 2, 2002
- Messages
- 3,672
By the sound of things, vertex shaders won't make any huge difference. Well, it's better than nothing and if you get no artifacts, why not I suppose. Even if it's a 0.1% increase, an increase for free is an increase for free. ^_^
Oh, and ManCannon, your issue could be caused by quite a large number of things. I note that you said it was after you unlocked the pipelines, have you tried closing them again to see what happens? Your problem is really getting a little beyond the scope of this thread if that's not true and I recommend you post in the video card forum itself (not just nvidia either IMO) unless relocking the pipelines fixes the issue. Thing is, jumpiness can be caused by anything from low memory (leaks maybe?) to a virus in your system and just a myriad of things in between.
Oh, and ManCannon, your issue could be caused by quite a large number of things. I note that you said it was after you unlocked the pipelines, have you tried closing them again to see what happens? Your problem is really getting a little beyond the scope of this thread if that's not true and I recommend you post in the video card forum itself (not just nvidia either IMO) unless relocking the pipelines fixes the issue. Thing is, jumpiness can be caused by anything from low memory (leaks maybe?) to a virus in your system and just a myriad of things in between.
I'm pretty sure it's the pipemod. I tested it some more and the higher I overclocked it the more frequently it happened. I don't really mind though. It only happens maybe once or twice a day (and I game a lot) and the performance gains I get from this more than make up for it. It's just odd that I don't get any artifacts or anything like that.
Do you think perhaps the card isn't getting enough power? I noticed someone on guru3d had voltmodded their 6800, and his overclock was similar to mine. He was running at 400/790 or so. My card tests fine at 400/800, and if I put it at that I get no artifacts and good performance, just the "hiccup" happens more often.
Do you think perhaps the card isn't getting enough power? I noticed someone on guru3d had voltmodded their 6800, and his overclock was similar to mine. He was running at 400/790 or so. My card tests fine at 400/800, and if I put it at that I get no artifacts and good performance, just the "hiccup" happens more often.
Chaballaman
[H]ard|Gawd
- Joined
- Jun 26, 2001
- Messages
- 1,515
i've never had single problem with 6800nu with 16, 6 pipes. gives me 5k on 3dmark05 and 12k on 3dmark03 so i can't complain. but just got 6800LE and its even better for the money. i unlocked extra 8,1 pipes, running 100%. max o/c so far is 380/870 but wasn't 100% stable. but i think i might be able to get almost 5k on 3dmark05 with 6800LE.
ive flashed it 6800nu bios but didn't unlock any extra pipes, seem to help o/c bit better with 6800nu bios tho..
but i think i might be able to get almost 5k on 3dmark05 with 6800LE. ive flashed it 6800nu bios but didn't unlock any extra pipes, seem to help o/c bit better with 6800nu bios tho..
Chaballaman said:i've never had single problem with 6800nu with 16, 6 pipes. gives me 5k on 3dmark05 and 12k on 3dmark03 so i can't complain. but just got 6800LE and its even better for the money. i unlocked extra 8,1 pipes, running 100%. max o/c so far is 380/870 but wasn't 100% stable.
ive flashed it 6800nu bios but didn't unlock any extra pipes, seem to help o/c bit better with 6800nu bios tho..
Thats because the pipes are hardwarelocked (only the Rivatuner softmod works) and the 6800nu´s have a higher standard voltage than the 6800le´s. What drivers are you using and what are the specs of your computer? 12k in 03 sounds incredible, almost unbelieveable for a 6800nu @ 16/6, unless you run it at least 410/950 or even more.
Nazo
2[H]4U
- Joined
- Apr 2, 2002
- Messages
- 3,672
I agree with this. 6800LE uses less power, so flashing the bios gives you a video card using more power, so it overclocks better (and gets warmer btw, so check that cooling.)karl76 said:
Don't know about drivers, but, he did put his specs in his signatures... I don't think 12K in '03 or near to 5K in '05 sound unbelievable for a 6800nu... Er, do you have signatures turned off? Take a look at mine. Ok, not 12 or 5, but, not THAT inconceivable that someone with a similar setup might just pull off those numbers. IMO, if I could only run the FSB at 200 instead of 180 I may well actually see those numbers. Believe it or not, unlocking the extra pixel pipelines makes an absolutely incredible difference on this card. Oh, and a nu isn't THAT different from a GT btw. It's not until the ultra that you start to see the true differences beyond a very little bit of overclocking. (It has an even higher voltage and the card has regulators and such to actually handle it, so it can oc even more.) In fact, I could run my core up as high as 425 with a decent voltage with all pipelines unlocked. 450 was stable in everything but Doom 3. Unfortunately, that voltage would tear up my card *sigh*
Nazo said:
I missed his sig, silly me, but yes his nu is running @412/931 and of course 16/6, plus his FSB is much higher than mine. I added my system to the sig, thanks for reminding me! Anyway, my old scores were 8300/3300 (with 66.93) with the 6800nu at stock settings as I wrote in my earlier post, I did overclock it now to 370/800 (@12/5) and managed 9300/4000 in 3Dmark03/05 (with 71.84), but you see I´m still far away from such high scores, especially in 3Dmark03. I know I could get a few more points with the 75.90 drivers, but definitly don´t feel comfortable to overclock my card any further, I want to keep it for a while.
I´m wondering if mainy my CPU is holding me back, I have a XP2400@2700 (166x13), or if its the settings that my 6800nu is running at of if its a combination of everything, including drivers.But thank you for your response, it made me think about this and thats always good!
Your 3dmark03 seems low to me karl. When I clock my card to score 4,000 in 3dmark05 it usually at least scorees 10,000 in 03. Shit, at stock with no OC I scored 9,100 in 03 with a 05 of 3,400. An increase of 600 points in 05 should net you around double that many points in 03.
Nazo
2[H]4U
- Joined
- Apr 2, 2002
- Messages
- 3,672
From what I understand, the Barton is officially CPU limiting -- even one running at > 2GHz+. Probably not bad though. Well, I for one see pretty good results in even the most modern games on mine -- just wish my memory wasn't holding me back because this CPU can quite easily pull off 2.5 and I'd definitely see a 5K or so score then. (I believe I originally tested with my memory desynced and the CPU set to 200x12.5 and, in fact, got around 5K but, I didn't save those results since it was just a test to see what would happen. With the memory desynced, games jump around so much it's painful.) Anyway, he did say probably. He probably overdid it a little -- I admit I find it hard to imagine a 6800LE hitting near 5K in '05, but, with a good system, you may see some pretty good numbers on a LE, especially if you can unlock it. Also, the flash definitely gave it more overclocking on that core for even better results I'm sure.
Stompya, I think it should be possible, yes. As I recall, there were two options I had to check for the pixel shaders and one for the vertext. If I'm remembering right, then I would presume you can select only one of the two instead. Guess you'll have to test both ways to figure out what is stable (assuming any are.) No harm in trying right?
Stompya, I think it should be possible, yes. As I recall, there were two options I had to check for the pixel shaders and one for the vertext. If I'm remembering right, then I would presume you can select only one of the two instead. Guess you'll have to test both ways to figure out what is stable (assuming any are.) No harm in trying right?
You can't just unlock a couple pipes, it's all or nothing. But as he just said, you can enable the vertex or pipes separately and see if one or the other will work on it's own. That's what I did. I unlocked both of them at the same time and thought I got a card that wouldn't unlock, but then I enabled the pipes and not the vertex and everything was fine.
ManCannon said:
I think 3dmark03 depends more on FSB than 3dmark05, its the only explanation that I have for my low 03 scores. I have an old Tbred that I´m running at 166x13, my stupid motherboard doesn´t like FSB200, the northbridge gets so hot that the whole system gets so unstable and locks up after a few seconds. I might try taking off the northbridge passive heatsink, mirror polish the buttom, and apply AS5 and active cool it. Or I just get a used Deluxe board with the higher revision (FSB200 works fine on them), get a Barton CPU, but that´d cost like $150, for that price (or slightly more) I could get an older (754) AMD64 incl. motherboard and have a faster CPU in general.
By the way, whats your FSB running at?
Nazo said:
Yes nforce2 motherboards definitly hate to have desynced FSBs, but one thing I don´t understand with your memory problem is that it actually is designed to run at least at FSB200 if its PC3200, right? And you have PC3200 Geil memory. Why didn´t you RMA it? Also I read somewhere that Geil memory likes to have some extra voltage to run stable/overclock better. I bet you tried out many settings, but I wanted to share with you what I read anyway.
Nazo
2[H]4U
- Joined
- Apr 2, 2002
- Messages
- 3,672
Yes. Yes it is PC3200. Unfortunately, it's YEARS past the RMA period. You see, my dad bought this memory before there WAS a DDR400. At the time, the only way to achieve 400MHz (200x2) was to overclock the heck out of your system, and Geil didn't have any that could OC that high, so they were never able to test them at DDR400 speeds. And, heck, when I first got this memory I was running a board and CPU that could only handle around 133 (ok, it oced to around 145 or so) so even then I didn't know. I'd say that it's well past all forms of warantee by now. As for the voltage, well, it's set as high as it goes in the bios. I think my options were 2.6 and 2.7, but, maybe they were 2.7 and 2.8. Either way, it won't go higher. I tried lower also, just to be sure it wasn't too much, and things where still unstable past 182 or so.
Problem is, I just couldn't bear to have less memory. I'm addicted to what it's like having 1GB. There's just always enough for everything, you know? I have a ton of things running, set them to use all the memory they want, and games still run smoothly. To get 1GB of PC3200 costs rather a painful amount still today. Cheapest I ever saw was around $100, and that was a brand I didn't trust which didn't even so much as have heatsinks on it (ok, I could always get some, but, they don't help THAT much.)
PS, about nForce 2 boards. I've gone through several I'm afraid to say, and, I've noticed that all the cheap ones (never bought an expensive one) have a weak SOUTHBRIDGE. Yes, that's right... Problem is, I keep trying to cool the southbridge better and end up doing it all wrong, so I make them more unstable. I think I put too much compound and, also, I should have used a bigger heatsink, but, I was afraid it would get in the way of my video card -- which is quite large IMO. The ASUS I used to have (-x, aka cheap) wouldn't even make it to 200 when I was done because I seriously screwed up and had some superglue on there adding insulation. (I think it started rebooting around 190 at the highest.) Mind you, the northbridge is as sensitive as it ever was, and I stuck my old Radeon Iceberq on there (little drilling, but, hey, it's larger than the normal chipset hsf they offer and the whole thing is a very nice high copper alloy.) Due to the southbridge thing, 200 FSB is unstable whenever it hits 31C or so (just reboots) and I can't OC beyond 200 at all hardly without reboots. (Think I got 202, but, what's the point? Even on the CPU that's only 25MHz and even for a Barton it's not THAT much difference...) I suspect that whenever your system simply reboots due to OCing or whatever, it may well be the southbridge limiting you. Of course, hard to be sure, it's possible for the northbridge to do that too and I suspect that despite the manufacturer's claims, not many make it to 200 in a room that isn't inside a refridgerator with passive cooling on the northbridge. I would, of course, recommend cooling the northbridge better (passive = ouch) for one. I would personally recommend the Vantec IceBerq ( http://www.newegg.com/app/ViewProductDesc.asp?description=35-110-103&depa=0 ) as these people seem to be rather serious. My IceBerq 4 (made for a video card, NOT a chipset, so don't get that) is true copper rather than being one of those stupid things where the company calls it copper, but, what it really is is a slightly copper colored paint of some sort and if you scratch it you immediately find out their so called 100% copper heatsink is 99% aluminum, 1% paint... Also, one neat factor is that when I first got the thing, I could see my reflection quite clearly in the bottom. No lapping required -- in fact, you can only make it worse with even fine grit sandpaper. If I were you, I'd get one of those and some thermal paste (here's arctic silver's paste: http://www.newegg.com/app/ViewProductDesc.asp?description=35-100-005&depa=0 ) and then put your northbridge heatsink on the southbridge. Just be careful you don't screw up like I have so many times and not get it on there right or something.
Problem is, I just couldn't bear to have less memory. I'm addicted to what it's like having 1GB. There's just always enough for everything, you know? I have a ton of things running, set them to use all the memory they want, and games still run smoothly. To get 1GB of PC3200 costs rather a painful amount still today. Cheapest I ever saw was around $100, and that was a brand I didn't trust which didn't even so much as have heatsinks on it (ok, I could always get some, but, they don't help THAT much.)
PS, about nForce 2 boards. I've gone through several I'm afraid to say, and, I've noticed that all the cheap ones (never bought an expensive one) have a weak SOUTHBRIDGE. Yes, that's right... Problem is, I keep trying to cool the southbridge better and end up doing it all wrong, so I make them more unstable. I think I put too much compound and, also, I should have used a bigger heatsink, but, I was afraid it would get in the way of my video card -- which is quite large IMO. The ASUS I used to have (-x, aka cheap) wouldn't even make it to 200 when I was done because I seriously screwed up and had some superglue on there adding insulation. (I think it started rebooting around 190 at the highest.) Mind you, the northbridge is as sensitive as it ever was, and I stuck my old Radeon Iceberq on there (little drilling, but, hey, it's larger than the normal chipset hsf they offer and the whole thing is a very nice high copper alloy.) Due to the southbridge thing, 200 FSB is unstable whenever it hits 31C or so (just reboots) and I can't OC beyond 200 at all hardly without reboots. (Think I got 202, but, what's the point? Even on the CPU that's only 25MHz and even for a Barton it's not THAT much difference...) I suspect that whenever your system simply reboots due to OCing or whatever, it may well be the southbridge limiting you. Of course, hard to be sure, it's possible for the northbridge to do that too and I suspect that despite the manufacturer's claims, not many make it to 200 in a room that isn't inside a refridgerator with passive cooling on the northbridge. I would, of course, recommend cooling the northbridge better (passive = ouch) for one. I would personally recommend the Vantec IceBerq ( http://www.newegg.com/app/ViewProductDesc.asp?description=35-110-103&depa=0 ) as these people seem to be rather serious. My IceBerq 4 (made for a video card, NOT a chipset, so don't get that) is true copper rather than being one of those stupid things where the company calls it copper, but, what it really is is a slightly copper colored paint of some sort and if you scratch it you immediately find out their so called 100% copper heatsink is 99% aluminum, 1% paint... Also, one neat factor is that when I first got the thing, I could see my reflection quite clearly in the bottom. No lapping required -- in fact, you can only make it worse with even fine grit sandpaper. If I were you, I'd get one of those and some thermal paste (here's arctic silver's paste: http://www.newegg.com/app/ViewProductDesc.asp?description=35-100-005&depa=0 ) and then put your northbridge heatsink on the southbridge. Just be careful you don't screw up like I have so many times and not get it on there right or something.
Nazo, first off all thank you for your detailed, informative and helpful response, I highly appreaciate it! Considering when your dad bought the Geil memory, it´s almost a surprise that it is working on your motherboard, since certain motherboards definitly don´t like all brands there are and that aren´t matured enough, which could be said with your memory. You could always try to sell it on ebay, I´m currently selling my old noname 2x256MB memory chips and the auction goes surprisingly well, so in the end I have paid maybe $60 for decent 1GB ram after selling the old memory.
The reason why I think the its northbridge is because the way it is getting unstable (heat unstablity type) and that it is bascially the only major component that is only passive cooled so far. My motherboard (with nforce2-SPP and NOT nforce2-400 or 400 Ultra) originally only supported FSB166 at most, Asus released a BIOS update later on that added FSB200, however the revision of the northbridge was the old one that already got pretty hot (but stable) at 166 but got unstable/too hot at 200. The southbridge on the other hand was already crazy hot at 133, when touching it with your finger you could literally fry it and considering what important components are stored in there, I active cooled it pretty early on. Again later revisions (2.0) of the same board with later stepping of the nforce2-SPP chip solved the heat issues of the north and southbridge and these versions are also (or suppose to be) as stable at FSB200 just like the nforce2-400 chips were in the first place. I definitly should have bought the "Deluxe" version (nforce2-400) in the first place, at least I learnt the lesson to get always the Deluxe/Ultra versions of motherboards if they offer better performance features.
Thanks again for your advice and great post!
The reason why I think the its northbridge is because the way it is getting unstable (heat unstablity type) and that it is bascially the only major component that is only passive cooled so far. My motherboard (with nforce2-SPP and NOT nforce2-400 or 400 Ultra) originally only supported FSB166 at most, Asus released a BIOS update later on that added FSB200, however the revision of the northbridge was the old one that already got pretty hot (but stable) at 166 but got unstable/too hot at 200. The southbridge on the other hand was already crazy hot at 133, when touching it with your finger you could literally fry it and considering what important components are stored in there, I active cooled it pretty early on. Again later revisions (2.0) of the same board with later stepping of the nforce2-SPP chip solved the heat issues of the north and southbridge and these versions are also (or suppose to be) as stable at FSB200 just like the nforce2-400 chips were in the first place. I definitly should have bought the "Deluxe" version (nforce2-400) in the first place, at least I learnt the lesson to get always the Deluxe/Ultra versions of motherboards if they offer better performance features.
Thanks again for your advice and great post!
Chaballaman
[H]ard|Gawd
- Joined
- Jun 26, 2001
- Messages
- 1,515
6800LE rocks! unlocked extra 8, 1 pipes with rivatuner, gives me slightly under 5K in 3DMark05.. (n i only paid $170ea for two, both unlocked fine)
6800LE is not much slower than the my o/ced 6800NU (16, 6 pipes, 410/930), which is the fastest 6800NU i've seen in here.
6800LE is not much slower than the my o/ced 6800NU (16, 6 pipes, 410/930), which is the fastest 6800NU i've seen in here.
biohazard_nz
Limp Gawd
- Joined
- Jan 2, 2005
- Messages
- 461
Chaballaman said:
Thats almost 6800GT clocks isn't it? (410/930)
Does yours have GDDR3 memory? Because if it did then there might be a possibility of flashing to a 6800GT
Kckazdude, could you link these instructions from the first post? Add http://www.hardforum.com/showthread.php?p=1027369748#post1027369748 to link directly to it.
Here's step by step instructions for unlocking the NV43:
1. start rivatuner and click on customize in the target adapter box.
2. Click the first icon, "Low-level system settings"
3. Click Install to install the nvstrap driver
4. Click the "allow enabling hardware masked units" checkbox, then
5. Click the customize button
6. Click to the left of whichever pixel pipeline quad is listed as "disabled" to enable it and click OK. Then click OK on the Low level system tweaks window
7. Click OK to reboot the system
8. Start rivatuner again to verify that all 8 pipelines are now enabled
Done.
One note: coming out of hibernation mode resets the pipeline masking, so a reboot is necessary to re-enable all 8 pipelines.
Here's step by step instructions for unlocking the NV43:
1. start rivatuner and click on customize in the target adapter box.
2. Click the first icon, "Low-level system settings"
3. Click Install to install the nvstrap driver
4. Click the "allow enabling hardware masked units" checkbox, then
5. Click the customize button
6. Click to the left of whichever pixel pipeline quad is listed as "disabled" to enable it and click OK. Then click OK on the Low level system tweaks window
7. Click OK to reboot the system
8. Start rivatuner again to verify that all 8 pipelines are now enabled
Done.
One note: coming out of hibernation mode resets the pipeline masking, so a reboot is necessary to re-enable all 8 pipelines.
Im also using a EVGA 6800NU with a bios flashed to 1.5volts.
All the pipes are unlocked 16/6 running the core at 435-440 100% stable with the memory around 870-890. Im able to hit 12900 in 3dmark2003 and Im currently in First place on 2005 with a P4 and 6800NU on the ORB Database. Been running this card for over 6 months cranked like this. Farcry, Doom3, Halflife2 all are max in game settings and 16x Anisotropic set in the vid card. Using riva tuner to unlock and overclock.
All the pipes are unlocked 16/6 running the core at 435-440 100% stable with the memory around 870-890. Im able to hit 12900 in 3dmark2003 and Im currently in First place on 2005 with a P4 and 6800NU on the ORB Database. Been running this card for over 6 months cranked like this. Farcry, Doom3, Halflife2 all are max in game settings and 16x Anisotropic set in the vid card. Using riva tuner to unlock and overclock.
Nazo
2[H]4U
- Joined
- Apr 2, 2002
- Messages
- 3,672
ManCannon said:
1.2 is stock 3D for a nu. Personally, I've been trying to get 1.3, but, Nibitor only lists 1.1, 1.2, 1.4, and 1.5 (see my thread "Effects of voltmodding on lifetime for my discussion about this whole mess.) Some people get good results on 1.4, but, it also tears up a number of cards, so I personally wouldn't recommend risking it (if you had that much money, why not just buy an ultra to begin with...) 1.5 is probably guaranteed to fry a huge majority of the cards, so I wouldn't count on being one of the lucky ones.
I have so far been unable to get 1.3 volts, however, when I ran 1.4 I saw 425 stable on my core, so, if you can find a way to get 1.3, I strongly recommend giving it a shot. Actually, at 1.4V, I was able to set 450, and the only time I ever saw an artifact was in Doom 3 (or was it locking up? I don't remember.) Bear in mind that many NUs have some of the pipes disabled for a reason and won't OC any more stable regardless of voltage once unlocked.
I actually just went back to 12 pipes and tried overclocking, and I got lockups with that too, so I don't think it's the pipemod doing it. I'm thinking the card probably needs more juice. So Nibitor only lets you go to 1.4 huh? Did you try that other one, Omniextremesomethingorother, or does that one only allow you to go to 1.4 as well?
EDIT: Nevermind, read your thread, seems you've tried Omni too. In my travels reading about this, it seems a few guys over at guru3d have had their NU's voltmodded and unlocked for a while and I haven't heard any complaints about their cards burning out. Still, I don't think it's worth the risk, but if people can pull of 1.5 and 1.4 volts for months at a time maybe we would be safe with 1.3. I wonder why that's not an option in any of these BIOS editors.
EDIT: Nevermind, read your thread, seems you've tried Omni too. In my travels reading about this, it seems a few guys over at guru3d have had their NU's voltmodded and unlocked for a while and I haven't heard any complaints about their cards burning out. Still, I don't think it's worth the risk, but if people can pull of 1.5 and 1.4 volts for months at a time maybe we would be safe with 1.3. I wonder why that's not an option in any of these BIOS editors.
Nazo
2[H]4U
- Joined
- Apr 2, 2002
- Messages
- 3,672
You know... 1.4 doesn't damage them for a month or so. You might try setting that and testing to see if it's stable, then going back to 1.2 until someone figures out a way to set 1.3. Just in case though, have the standard PCI video card and bootdisk and all.
That way you can find out if your problem really would be solved by a better voltage.
That way you can find out if your problem really would be solved by a better voltage.
Hi this is my first time with over clocking video card and was just wondering i have hte pny 6800. i followed the whole instruction i the start and got the pixel and vertex okey. When i tried overclocking it wouldnt budge even on the stock settings the test would fail. then i tried disabling the pixel and and tried changing the core and mem clock got it to 390/820 working okey. then i tried to put back and pixel thing and it still works. so does this mean its okey or did i just force it into somthing that will cause it to break? will the increas in clock cause it to heat up more? how do you test for stability? would greatly appreciate any help thanks