I got 00000174:000037a0.
I have no idea how to translate that.
I'm learning...
37a is 890, thanks to a google search
so 890/1023= 86.99%
That sound right?
It looks like that is close to mine but a little higher, mine being 883/1023= 86.3%
Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
I got 00000174:000037a0.
I have no idea how to translate that.
Running heaven, looking at statue
volts @1218 trixx = 1.115 in gpuz 1180mhz will crash at 87 degrees (fan 57%)
volts @1193 trixx = 1.152 in gpuz 1180mhz will crash at 90 degrees (fan 54%)
volts @1170 trixx = 1.068 in gpuz 1180mhz will ---- artifact at 95 degrees (fan50%) - took longer to heat
volts @1170trixx
1240Mhz artifacts occur @71 degrees
1230Mhz artifacts occur @73 degrees
1220Mhz artifacts occur @75 degrees
1210Mhz artifacts occur @80 degrees
1180Mhz artifacts occur @95 degrees
so the main thing i notice before crash, the vddc and Vddc current change drastically
when the card starts out VDDC current would be around 140 amps
when the card heats up the Vddc keeps going up, eventually around 190, 1 recording log from gpuz even captured an odd 250 (followed by 2x 180 then crash)
The load is always the same, standing still looking at statue, seems once the temps raise the voltage control needs to be cooler, raising the voltage only hindered my OC efforts, resulting in crashes at lower temps
A) volts x amps + high temp = Vddc amps out of control = crashes
B) lower volts x amps = higher temps before vddc out of control
C) higher fanspeed = insane warhammer behind the computer
oc tools,
CCC, Sapphire Trixx, left mouse button
GPU-z 0.5.8 (new release with full volt monitors for 7xxx)
I saw you post in my thread over at Guru3D (Dr. K6 over there). Unwinder says there's no Vdroop, just the sensors reading incorrectly. The newest MSI AB adjusts for this. I believe what you're seeing above is a temperature wall. Temperature increases the resistance of the IC, hence the amps go up in order to maintain the voltage. However, I believe at some point the OCP kicks in and shuts down the card. Keep the card cool and it will clock better. At ~45C, I hit 1340MHz on the core with 1.3V.NICE post.
With all the people QQ'ing about their cards being defective because they cant reach 1125, it's finally nice to see a very helpful and insightful post again.
I can verify your findings directly.
My 1175mv (stock) card is rock stable at 1150 mhz, and 1700 Ram still at 1175 mv. I tested it in Gputool, and no artifacts up to 86C, where I stopped the test.
However at 1200 mhz, it seems that the artifacts start happening around 76C.
What's weird is, it doesn't seem to matter what the voltage is.I tried 1.215v (which was about 1.13 under load) up to 1.250v (which was about 1.16v under load), and all it did was raise the artifact level by 1 total C....basically it didnt seem to matter at all.
I do know that these GPUtool artifacts don't really tell the whole story, since Battlefield 3 is fully stable at 1.231v and 1200 mhz, but I'll see weird artifact lines on the planes in chase cam view at 1.21'ish volts. It seems like in games at least, raising the voltage does have some effect, so it's really hard to say what is causing the artifacts in gputool and what is causing the artifacts in games.
Similar results on the VDDC current. Saw that doing a furmark test at 1215 mhz and 1.25v, watched the amps shoot up slowly to around 250 (!) as the card reached 82C, and then the card itself *shut down*. The screen turned black and the fan speed went way down and the system was locked hard. That black screen videocard crarsh didn't happen at 1200 mhz but artifacts appeared at 86C.
I think we need more time to find out what's really going on here. And I don't even know if a vdroop mod would do anything or not for this...
Furmark is useless now since both AMD and NVIDIA have protection in their cards against power viruses. The program can't max out the card, therefore you can't produce the full load needed to test for stability. You need to test your overclocks in games. Games that run very high FPS (older source games, like Left 4 Dead are awesome here), work the best IMO. I've found running Crysis 2 benches works well too.Got to 1125, ran it through Furmark didn't see any artifacts or isntability. Then I fire up Skyrim and an insta-crash (of the display driver stopped responding type). Is Skyrim super sensitive to OCs or something?
Still debating watercooling, so just want to make sure I understand this correctly:
Since I have a 1.050V card, this gives me less headroom regarding how high I can raise the voltage past 1.175V before OVP/OCP kicks in. Is that right?
Still, would be nice to get my temps to where they're not occasionally hitting 89C @ 1.175V in normal game usage. I could use a custom fan profile, but the fan already gets pretty loud.
Even if you have to back them down a bit, I think you have a better card than I do. At 1.175V everything looked fine at 1175 (core) until I noticed some serious artifacting in BF3 after a while, and then discovered my system locks up in Anno 2070 at anything past 1100. So I'm curious if getting my temps down will help.I'm definitely getting a waterblock as soon as I can afford one. I hate the sound of this fan!! I don't know for sure about your OC headroom, but my card (1.05v) is stable in 3Dmark11 running 1220/[email protected] thus far (haven't tested this setting in games yet; I expect them to have to go back a little) using a custom fan curve in AB (temps around 68-70*C with fans maxing around 60-70%). It sounds like a plane taking off; my wife makes fun of me right now because of it, and that's unacceptable![]()
Even if you have to back them down a bit, I think you have a better card than I do. At 1.175V everything looked fine at 1175 (core) until I noticed some serious artifacting in BF3 after a while, and then discovered my system locks up in Anno 2070 at anything past 1100. So I'm curious if getting my temps down will help.
My fan gets pretty loud on auto as it is, and I'm usually not one to complain about noise. The default custom fan curve that's automatically set in Afterburner will keep it cool at around 74C, but that's reminiscent of of my old Vantec 92mm Tornado (which was too much even for me). I could mess around with it, but I'm not sure what the best settings are, and I'm not entirely sure how much more noise I want to deal with on a regular basis.
So, the whole watercooling thing is pretty tempting. Still, there are a lot of other things I could use that money for. Decisions...
This is kinda weird. I just reinstalled Windows and now MSI Afterburner is limiting my memory to 1790 MHz.
Only thing I can think of that is different is I downloaded the latest driver from AMD, RC11.
I installed TriXX which allows me to overclock to 1825 again, but then it doesn't have the monitoring window that Afterburner does.
I saw you post in my thread over at Guru3D (Dr. K6 over there). Unwinder says there's no Vdroop, just the sensors reading incorrectly. The newest MSI AB adjusts for this. I believe what you're seeing above is a temperature wall. Temperature increases the resistance of the IC, hence the amps go up in order to maintain the voltage. However, I believe at some point the OCP kicks in and shuts down the card. Keep the card cool and it will clock better. At ~45C, I hit 1340MHz on the core with 1.3V.
Furmark is useless now since both AMD and NVIDIA have protection in their cards against power viruses. The program can't max out the card, therefore you can't produce the full load needed to test for stability. You need to test your overclocks in games. Games that run very high FPS (older source games, like Left 4 Dead are awesome here), work the best IMO. I've found running Crysis 2 benches works well too.
I saw you post in my thread over at Guru3D (Dr. K6 over there). Unwinder says there's no Vdroop, just the sensors reading incorrectly. The newest MSI AB adjusts for this. I believe what you're seeing above is a temperature wall. Temperature increases the resistance of the IC, hence the amps go up in order to maintain the voltage. However, I believe at some point the OCP kicks in and shuts down the card. Keep the card cool and it will clock better. At ~45C, I hit 1340MHz on the core with 1.3V.
Dear K6,
Saying 1.3 volts and 45c is nice. BUT, what would be your hard and fast rule for NEW overclockers running on STOCK cooling?
Over 70% fan is too loud, my stable overclock are best acheived at stock volts, raising the volts for my stock cooler helps nothing. We have plenty of people buying these cards, they raise volts because thats what others have said, CRASH, then they want to RMA their cards because they wont oc to 1300mhz...
edit : also how does the temp wall affect artifacts? I understand the higher Amps at higher temps /crash, but why does it artifact at the higher temperature?
Furmark is useless now since both AMD and NVIDIA have protection in their cards against power viruses. The program can't max out the card, therefore you can't produce the full load needed to test for stability. You need to test your overclocks in games. Games that run very high FPS (older source games, like Left 4 Dead are awesome here), work the best IMO. I've found running Crysis 2 benches works well too.
Furmark is useless now since both AMD and NVIDIA have protection in their cards against power viruses. The program can't max out the card, therefore you can't produce the full load needed to test for stability. You need to test your overclocks in games. Games that run very high FPS (older source games, like Left 4 Dead are awesome here), work the best IMO. I've found running Crysis 2 benches works well too.
I wouldn't trust Furmark anyway because, as I said above, both AMD and NVIDIA specifically code their drivers to kill power virus type programs. It was my understanding that Unwinder changed the coding simply to show what people wanted to see, not what the sensors actually read. It's my understanding that the sensors read as a percentage of what the current is, and use VID (1.175V) as a max, and therefore a percentage of that in relation. If I'm wrong that's fine, but my card is stable as can be with no issues overclocking. Also note that when I go from stock to 1.3V applied, my power consumption goes up some 100W, so it's definitely taking more voltage. The difference here is I'm on water cooling, that's why I think it's a cooling issue.Unwinder didn't say that. He said he's adding the "Target" voltage back as default, due to too many noobs crying about the voltage reading incorrectly. That's what he said.
There is very definitely vdroop. You can use the latest version of GPU-Z, and notice weird stuff happening with the volts and amps especially in furmark...really bizarre..but that's how it is.
This doesn't change anything for me, I'm still reading the same (1.17 @ 1.3V applied, typically 1.12V load, and 0.85V idle), and my voltage never fluctuates or changes, in GPUZ or Afterburner.You can unlock the real measured voltage iin AB (to make it like beta 10) by editing the profiles folder after installing AB and rebooting.
Put
VDDC_CHL8228_VIDReadback = 0
MVDDC_CHL8228_VIDReadback = 0
in the [settings] under profiles and the VEN... .cfg profile, (not in the main msiafterburner.cfg file).
The first rule of overclocking is there are no hard/fast settingsDear K6,
Saying 1.3 volts and 45c is nice. BUT, what would be your hard and fast rule for NEW overclockers running on STOCK cooling?
Over 70% fan is too loud, my stable overclock are best acheived at stock volts, raising the volts for my stock cooler helps nothing. We have plenty of people buying these cards, they raise volts because thats what others have said, CRASH, then they want to RMA their cards because they wont oc to 1300mhz...
edit : also how does the temp wall affect artifacts? I understand the higher Amps at higher temps /crash, but why does it artifact at the higher temperature?
I use 25MHz increments, but yeah, my score actually increases until the RAM flat out crashes. Even if ECC is kicking in earlier, my FPS and score keep going higher. People with high RAM clocks are probably overvolting the vRAM to 1.7V (stock is 1.6V). At stock, my vRAM tops out at 1625MHz, but with 1.7V I can get 1775MHz stably. However, I don't recommend overvolting the RAM for very long, as vRAM has typically been very sensitive to voltage and can get killed quickly. Unless of course that's changed with these cards, but I'm not going to be the one to test itI find it a little strange everyone is saying max ram OC's
running 3d mark 11 i can max out RAM too......
but over 1450ram, the score goes down, so i set to 1420.
Has everyone been testing their 3dm11 scores with 10mhz increments till score goes down? I know it takes time but your 2mega jillion hertz (RAM) overclocks dont seem right to me....
Okay, I had my card crash at 1175 and 1125...so I'm jumping back down to stockI wonder why my system can't handle this
same
either that, or final fantasy VII
I find it a little strange everyone is saying max ram OC's
running 3d mark 11 i can max out RAM too......
but over 1450ram, the score goes down, so i set to 1420.
Has everyone been testing their 3dm11 scores with 10mhz increments till score goes down? I know it takes time but your 2mega jillion hertz (RAM) overclocks dont seem right to me....
^^
I was having problems overclock my first time through, and found out it's because I had Overdrive enabled while using Trixx to overclock. I am currently pulling 1.2/1600 at stock voltage and the card hanging around 70c. I'm testing it in SWTOR maxed out at 5760x1080 and so far no problems and no artifacts.
I was using gpuz 5.7. IT reported my sapphire 7970 to have 1.175 volts (while battlefield 3 is playing in window mode). I was using the drivers on the sapphire cd. I did get some minor artifacts, but I was able to play with all 3 sliders maxed (in the ati overdrive tool). I tried the 1/9 drivers; I can no longer do the 3 sliders maxed (without bf3 crashing to the desktop randomly). I can do like 1100/1575. I reformatted since then however, and I am using the latest 7970 drivers (came out a couple of days ago). GPUZ also released a new version, and I am using it. This gpuz is showing 1.07 voltage, with bf3 in window mode. SO WTH?
MY stock volts was apparently 1.175, but now it is 1.07 (under load)?
Old gpuz didn't monitor voltage correctly on the 7970. The new 0.5.8 does monitor correctly. I originally thought my voltage was 1175mv but it was actually 1050mv.
so thats why the max 3 sliders are giving artifacts. I will have to up the voltage to 1.175. So 1.175 is supposed to be the default voltage (as in perfectly safe. Like 100% safe.)? I know some cards come with 1.05-1.175. I am not sure why this is the case. Why can't they all by 1.175. I guess I ain't a hardware technician, that makes silicon shit. So I would not know the reason.
Yeah idd i wonder to that if you put 1.175 on 1.05v or 1.125v card will be bad. I hope someone can explain me if it is. I think personally all cards are same so all should be safe 1.175 if im wrong please say so![]()
... I guess I ain't a hardware technician, that makes silicon shit. So I would not know the reason.
so thats why the max 3 sliders are giving artifacts. I will have to up the voltage to 1.175. So 1.175 is supposed to be the default voltage (as in perfectly safe. Like 100% safe.)? I know some cards come with 1.05-1.175. I am not sure why this is the case. Why can't they all by 1.175. I guess I ain't a hardware technician, that makes silicon shit. So I would not know the reason.
If your card is a 1.05v card, then you shouldn't need to go as high as 1.175v to max CCC. I think I maxed CCC with 1.1v...
![]()
If anyone is interested I made a quick video of my nice air-cooled 1300Mhz 7970 overclock. I cannot believe how well the stock cooler holds up under this load!
http://www.youtube.com/watch?v=KRh3P3GJZWI&feature=youtu.be
Anyone else reach 1300Mhz on air?
If anyone is interested I made a quick video of my nice air-cooled 1300Mhz 7970 overclock. I cannot believe how well the stock cooler holds up under this load!
http://www.youtube.com/watch?v=KRh3P3GJZWI&feature=youtu.be
Anyone else reach 1300Mhz on air?
Well if i put alot of volts on it i have a feeling i can do even higher. But why would idont want to run my gpu at 1.3v. At 1.12v between 1.14v load i can run 1220 so maybe 1300 should be doable at 1.2v
I'll try when I get my cards under water. 70%+ fan is just too loud for me, and 1250+ overclocks require that on my end![]()