Thanks guys for the info.
I wouldn't be too concerned about the memory not clocking as high as you see some of us. I'm pretty certain that the memory timings on the GTX cards are tighter than the GT cards, and thats why you see such high clocks on the memory on GT cards. Its also a different memory chip from samsung, which is faster to begin with. As soon as Nibitor is updated to allow memory timing changes i'm going to play with them and see what works best.Originally Posted by Jii
Your scores DO seem kinda low though, but i'm pretty sure its not the memory clocks on the card that are the cause.
Marlon
DFI NF4 Ultra-D 0406BTA Bios
Opteron 165 CCBWE 0550 UPMW (naked)
2.7 GHz at 1.58v
2X512 Gskill Extreme DDR600 2.5-4-4-8 1T 300mhz
EVGA 7900GT 584/1800 (for now)
AquaXtreme 50Z, MP-05 PRO LE, MP-1 GPU Block
'77 Bonneville HC with two 120X38mm Sanyo Denki
Antec Neo HE 430
Opteron 146 CACJE 0551 2.8 GHz at 1.55v
ASRock 939 Dual SATA2
I found something i the BIOS called "PEG link mode", and the description of it was somethink like "... performance of the PCI Express card". Is this somethink I can tune to get a little more boost?
Finally, the 10k barrier was reached.
I had to use powerstrip to make these clocks. Coolbits just freezed the screen with core-clocks higher than 580.
Is that stock?! Thats a nice core if it is! Hey just a question, I just got my gt in today and the stock cooler is LOUD!!! I booted up the system for the first time a shut it down immediately because I heard a loud wooshing noise, after some searches I discovered its the gpu's fan! Lol, I have gotten way to used to water-cooling! Even with my high-end setup and loud rad fans this stock cooler is easily twice as loud as anything in my system. O well, gives me an excuse to strap my maze4 on it....
New Build Under Way...
Server:
X2 3800+ 0601 UPMW 2.95ghz w/ 1.48v cooled by zalman 7700 Cu
EVGA 7900gt CO Superclocked (725/875) cooled by modded AMD64 hsf
DFI NF4 Ultra-D
2 x 512mb TeamGroup pc4000 UTT BH-5
OCZ 520W Powerstream
get that baby going and let us know how it does.
X2 3800+ - IHS @ 2950 1.424v, 3000 1.488v
LDBHE 0601 TPMW "N" cooled w/ aquagate mini R120
eVGA 7900gt vanilla @ 625/1700 1.3vmod
DFI LanpartyUT nF4 Ultra-D (bios 702-1)
2 gb g.skill HZ 3:4:3:6 @ 250
Everything is stock, no mem or gpu mods. Stock-cooler to.
All straight out of the box, stock heatsink fan and NO volt mods.
![]()
--------------------------------------
1st eVGA 7900GT C0 Superclocked
- bios 5.71.22.14.04 dated March 9, 2006 see pic here. I dumped the bios and uploaded it here if anyone wants it.
- stock speeds = 550/1580
- Samsung BC14 1.4ns 610 (week 10, 2006) memory chips
- manages 590-595mhz on core and 1860 on mem fine when load temps are 63-66C which are close to same load temps at stock 550/1580.
- it managed 600/1860 when ambient temps dropped allowing 58-61C load temps but when load temp rises past 62+ it can't do 600/1860
--------------------------------------
2nd eVGA 7900GT C0 Superclocked
- bios 5.71.22.14.04
- stock speeds = 550/1580
- Samsung BC14 1.4ns 610 (week 10, 2006) memory chips
- doesn't seem as good as first card looping 3dmark05 already hit a limit on core - not passing at 590(610)mhz so far with 63-65C load temps
- limit of this card on the core seems to be 580-583mhz core and 1840mhz on memory with 62-65C load temps which are close to same load temps at stock 550/1580
--------------------------------------
3rd eVGA 7900GT C0 Superclocked
- bios 5.71.22.14.04
- stock speeds = 550/1580
- Samsung BC14 1.4ns 610 (week 10, 2006) memory chips
- has consecutive serial number with the 2nd card
- lowest load temps of the three eVGA 7900GT CO SC cards, @550/1580 load temp is 61-62C, idle temps around same 44-46C
- this card has similar core clocks to that of 2nd card but slightly better on core, around 585-595mhz core on this card and 1880mhz on memory
Looks like i'll be matching 1st and 3rd cards for SLI fun![]()
---
3D05 11743
3D06 6380
XFX 7900GTX 512 690/1750 clocked @ 725 / 1880 on stock cooling / no mods
Last edited by Raphie; 04-14-2006 at 12:41 AM.
Case: Thermaltake SHARK Black Case | PSU: TAGAN TG580-U15 | Mainboard: DFI Lanparty UT NF4 Ultra-D | Processor: AMD Athlon 64 X2 4400+ | Processor Cooling: Zalman CNPS9500 Led + Arctic S5 | Memory: OCZ EL DDR PC-4000 Dual Channel Gold 2x 1024 @ 3-4-3-8) | Storage: Maxtor DiamondMax 9 S-ATA RAID Striped (2x 120GB) + Hitachi Deskstar 7K80 SATA-II (80GB) | GPU: XFX Nvidia 7900GTX 512MB Extreme Edition DDR3 DUAL DVI | Display: Samsung 19" 930BF 4ms DVI
10 * 250 = 2500Mhz | Vcore 1.450 | Mem 2.6v 1:1 = DDR500 | 3Dmark06 6348
ELITEWARRIORS: A Formidable Fighting Force
hi i have a 7900gt clocks 569 core(570 will give srtifacts with atitool) and about 456 for the ram... temps are good too(46 highest 34 lowest)do u guys suggest the conductive ink volt mod?
will i gain a lot?
Last edited by adz; 04-14-2006 at 01:44 AM.
Try the 1.3vmod, the stockcooler can handle that vmod without problems.Originally Posted by adz
Agreed, 1.3v should be no worries for stock cooler. Saw temps in the 70's last night with 1.44v and 680 core on stock cooled GT, so things might get a bit toasty above 1.3vOriginally Posted by Zyprex^^
![]()
.:. X2 @ 2.9 .:. E6600 L628 @ 4.73 .:. AB9 @ 512 FSB .:. E6300 @ 100% OC Stable
U.K. overclockers: Represent your country, bench for EP-UK! Unite and show the world there IS a U.K. scene!
More bench action![]()
Still stock-cooler and stock volts.
![]()
Thought i'll pull this one out of the hat and no it ain't 3dmark05 score
@590(610)/1827 = 9,683
http://service.futuremark.com/compare?3dm06=245316
![]()
---
So..
http://www.xtremesystems.org/forums/...&postcount=679
http://www.xtremesystems.org/forums/...&postcount=711
http://www.xtremesystems.org/forums/...&postcount=713
....where I'm talking about possible chip-memory/PCB issues....
My old posts about G71 over 700Mhz follow to the “freeze” issues: PC Perspective
Right after these news - many sites starts panic against GTX “freeze” issues. It was premature panic since now we have words from experienced peoples about the nature of these issues. After more detailed investigation of posts on different forums - I have conclusion:
1) First of all: NVIDIA says that some cards having isues when working on 700/1800 and higher- it's related to problems in communication between core and memory. I suppose it's due to different driving oscillator options for ROP, Geometry and memory.
2) The first words about pre-overcklocking from BFG, XFX, EVGA - the main reason - it's not true since many peoples says that even if they sets default clocks on these cards(pre-overclocked and with "freeze" issues) - it doesn't helps - the issue is still present. These cards have some defect element in their PCB - it can be bad chip or bad element on PCB(just try to remember issues with BlackScreen on 6800 cards - even on default clocks - it was releated to bad capacitors). These cards were RMA'ed and new cards were delivered to the owners - no issues on new cards.
3) The main number of posts with "freeze" issues were from EVGA owners. The root case was in bios since they were not able even to install default drivers from EVGA CD. Also, after new bios flashing some users still having problems with "freeze" issues - it's due to 2).
4) There are a few posts post in "freeze" threads about high temperature. Some GTX cards have bad or not correct installed cooler. Right after installing these cars from the box and playing games - the temp was more than 90C. RMA'ed - no issues now.
5) Some posts with "freeze" issus were solved via bios tuning. It was some kind of platform issues. (i.e decreasing RAM timings)
6) Some posts with "freeze" issues were related to overheat issues.
7) To avoid "freeze" issues - some time ago XFX and EVGA decrease their delta in the bios to +40. I suppose that first delta +50 was not correctly tested and it's hasty conclusion from Nvida.
8) The first reason of "freeze" issues due to OC - over OC: highly overclocked GPU - more than 720MHz is not stable due to clocks drops during high 3D load. It's all on default voltage for GTX and 1.55V for GT. It can be solved by increasing volts. Also, many 7900GT cards limited by 700-750MHz due to simple PCB and gpu limits.
9) The second reason of "freeze" issues due to OC - PCIe issues: proved that crads with more than 700-720 experienced problems with PCIe freq. I suppose it's due to synchro reasons between GPU driven clocks and PCIe. It can be solved by carefully PCIe tuning.
I suppose that new drivers - trying to restore 3D clocks after failure since if you check clocks after 3DMark run(with freezing in some subtests) - it will be 3D clocks. Also, if you check scores in bad subtests - you will see that scores in these subtests related to 2D or low-power 3D clocks.
So... If you have problems with "freeze" or PC shutdown during games on 650/1600 - forward this card to manufacture: it's real defect issue. If you have problems on clocks higher than 700MHz - check your temp and volts. In common case: 600-610MHz it's limit for 1.2V; 690-700MHz it's limit for 1.4V; 720-750MHz it's limit for 1.55V. If you volts and temp are correct - check you PCIe freq. In common case: you need to increase PCIe freq: it should be done carefully -> +1MHz PCIe freq then try to find stable OC for your card -> then repeat it: +1MHz PCIe freq and so on...
Hope this helps you to understand and determine your problems.
ps: sorry for my bad engl..."from Russia with love" (c)....
Last edited by Nilson; 04-14-2006 at 11:47 AM.
.inimitable.
I just wanna throw in my results with the 1.5v mod on the core and the VF900-Cu and my 146 opty @ aprox 2971-3040mhz - I had a real hard time getting to the 12k mark, but i did itcard clocks is 722/953 which where my max, before i got lockups or lower score.
05: http://service.futuremark.com/compare?3dm05=1955471 : 12003
03: http://service.futuremark.com/compare?2k3=4738130 : 24020
I removed the 1.5v mod afterwards, because I think my cpu is a bottleneck even at 3ghz with a 650mhz on the core
Iam happy with my system though, for now
/Grau
My card "freeze" when I get over 610/1850, no mods done. Is this a "real defect issue" or is it just that the card ain't got enough volts?
It's limit for 1.2V.Originally Posted by Zyprex^^
GOESA, yep. Nice decision. It's quite correct.
.inimitable.
I'm so glad. With 1.55 V mod on water
711/900 stable
YAY!![]()
------------------Pushing the delta between idle and load, to the limit------------------
KoyoThread
7900GT 705/2040, DFI Ultra D AE0, Opty 146 3.10 Ghz, MP01 +MP05 . CPU: 19/21 GPU:24/28Fan/Shround in progress..
456 for the ram?Originally Posted by adz
46 highest (aka load temp for GPU)!?
34 lowest?
And you get 569Mhz??
- what is the real Vram speed.. check in coolbits (nVidia driver)
- how did you measure these temperatures? Are they with stock heatsink? Please dont tell me you ran benchmark, then opened nVidia driver panel, and THEN wrote down that number as load temperature. You have to use rivatuner so it can log temperatures WHILE the benchmark is running (alt-tab windows, or checking after benchmark finished will give you lower numbers because GPU already cooled abit).
- once again, many people like you, with stock heatsink want to do voltmod. How much did you pay for this video card? be honest.. Do you want to break it because you are too cheap to spend $20 on after market video card cooler? Its the same as the conductive-ink problem: some of you people spend $300+ on this video card, and then keep bothering everybody about if pencil will work, because you wont spend $5 on some conductive-ink.
So to answer your question.. go ahead, and try 1.3V, even 1.55V, but dont complain if your overclock is limited, or if you burn up your card.
24/7: A64 3000+ (\_/) @2.4Ghz, 1.4V
1 GB OCZ Gold (='.'=) 240 2-2-2-5
Giga-byte NF3 (")_(") K8NSC-939
XFX 6800 16/6 NV5 @420/936, 1.33V
Please pay attention to what settings others are using! You are probably incorrectly comparing your overclocking results to those with 7900GT and 1.4/1.5V mod, or GTX owners.Originally Posted by Zyprex^^
You are using only stock heatsink and only stock 1.2V. Your result is VERY VERY GOOD, and there is nothing wrong with your card. There is no "real defect issue".
Nilson, I think you are setting unrealistic expectations for people with the voltage scaling clock speeds you posted. Afterall, you say 610 is limit for 1.2V and eva2000 could barely get 590 on each of his 3 cards.
I think the following are much more reasonable values (assuming adequate cooling):
1.2V: 580
1.4V: 640
1.5V: 680
Ofcourse even though these are "reasonable" values, they are by no means "conservative", so dont starting posting complaing your card didnt reach 680, or something like that. I think a lot of folks here (especially with low post counts) are forgeting that a big factor in overclocking is "luck".. each GPU is different... maybe a few peopel can get a really good result, but many wont/cant.
24/7: A64 3000+ (\_/) @2.4Ghz, 1.4V
1 GB OCZ Gold (='.'=) 240 2-2-2-5
Giga-byte NF3 (")_(") K8NSC-939
XFX 6800 16/6 NV5 @420/936, 1.33V
yeah i think with 7900GT stock 1.2v you're doing exceptionally well on stock cooler if you can hit past 575+ mhz on the core
---
Here my Results Single 7900GTX on Air no Mods
2006 6508 http://service.futuremark.com/compare?3dm06=220572
2005 12483 http://service.futuremark.com/compare?3dm05=1934256
2003 24965 http://service.futuremark.com/compare?2k3=4730932
2001 39221 http://service.futuremark.com/compare?2k1=8948711
Aquamark 124952
i use a zalman vf700 coolerOriginally Posted by ***Deimos***
, temperatures are taken from atitool running in background (and it records highest lowest and mean temp)and ram is set also with atitool since coolbits(i don't know why) stops @ 800mhz ... so i have to use atitool. for the temperatures i used atitool instead of rivatuner because i thought rivatuner whould not recognize it at all... but i'll try it!i'll look for conductive ink(if i find it,i'll look in some hardware store) anyway tnx 4 your suggestions!
Bookmarks