Discussion in 'MSI AfterBurner Application Development Forum' started by msi-afterburner, Jan 6, 2012.
Gentlemen, including Unwinder .. may I remind you that this is an English forum.
Well I have in fact read that section which you would know had you bothered to ask instead of being dismissive and I am well aware of The I2C collision issue when running multiple tools but this is not what is happening in my case as the only monitoring tool that I run is Afterburner I did my research and conferred with alot of other people before I came here and reported the issue.
If it would be helpful to you I can post screen captures of the voltage spikes as well as my process manager to show that no other hardware monitoring programs are running.
The only think able to cause a spike is an I2C collision. Period. Discussion closed.
But I am only running one voltage monitoring program so what could possibly cause a collision?
You can clearly see the voltage spike and the I have no other monitoring software of any kind running.
Not trying to be mean I appriciate all the hard work you put into this application just trying to get to the bottom of this issue.
I have nothing to add to the things stated above, nothing but I2C bus access collision can cause it. And I'm not 24/7 answering machine, especially for nailig down and solving issues with ASUS cards, sorry.
1.No one asked you to be an answering machine. this is a afterburner forum where people come to discuss afterburner and issues they are having with it I'm simply pointing out an issue.
2. it's not just an ASUS card issue these are reference cards. Which mean they all use the same part regardless of which company made them, which means we are talking about an every card issue. If you would like I can replicate the same issue on my room mates MSI 7970.
mhmm, did i missread something or do we realy get Vram Voltage controle for Retail HD7970 ?
@Thorok Try reinstalling Windows. Only then you can be sure that nothing else can interfere! Have you even tried in safe mode ?
Yes, for both reference design 7970 and 7950.
Parteeey! Can't wait for Beta 11 to hit the market (if you can say it like this )
When will it come ? Guess when MSI gives the OK ? ^^ YO MSI! What are you're plans/timetable for launching Beta 11 ?
Thx a lot
Soon. I've submited it to MSI yesterday so it can be expected every minute.
DUT localization beta10
don't know if I'm to late since you are releasing 11 ?
11 will be faster sorry for it !
uhm, just wondering, but will beta 11 show Vram Temp, VRM temp ? Saw a screenshot some awfull looking tool that was called gpu tweak or such, which had these.
My mistake, I thought it was another issue of hysteria that beta 11 already addresses.
No, it will not for reference design cards. Once again, VRMs are not the devices intended for safe frequent polling. VRM tems will be only available on r7970 Lightning series having a dedicated sensor for it.
Enable (or disable) 64 bit High Precision Event Timer and see if you still have the problem.
A few interesting observations on Thaiti GPUs. First, it looks like similar to Fermi, Thaiti GPUs are binned and there are 4 types of Tahiti GPUs with different ASIC leakage level. So depending on it there are 4 possible default 3D voltages: 1.17500V, 1.11250V, 1.05000V and 1.02500V.
Currently all voltage control tools use hardcoded default voltage (1.175V) for 7970, so pressing "Reset"/"Defaults" buttons in any voltage control tool will put any GPU to 1.175V regardless of leakage level.
It looks like I have located the fuses identifying GPU leakage level, but I need more statistics on it. So if you have 7970 please run MSI Afterburner with the following command line switch:
Then post the result here, and don't forget to mention your default voltage as well.
Ref vlt is 1174 mV
Mine in 1175mV as well with value of 000023e0. As far as I can understand BIOS code, this value is ASIC leakage level fused during ASIC binning and manufacturing. An ideal GPU with no leakage current is 00003ff0, lower values indicate higher leakage and higher default voltage.
Mine is 00000174 and 00002f50
default voltage is 1175 mv.
Am I correct to assume that:
1) Higher leakage gpu's run hotter (just like "leaky" Intel core i7's will run hotter?)
2) Lower leakage (lower default vid, e.g. 1050 mv) GPU's will NOT overclock to the max CCC limits (some people with 1.05v GPU's said they couldn't come close to the CCC limits without artifacting, until they changed the voltage to 1175 mv), but if vcore is raised to 1175 mv, possibly these GPU's will overclock higher at a given voltage than our 1175 mv chips?
What's interesting is that "leaky" Intel CPU's usually ran hotter *BUT* also had a lower default VID, AND tended to overclock better (as long as you could cool them). I had a "cool" 2600k that ran at least 5C cooler than my current one, had a high (1.2375v) default VID at 3.4 ghz, and needed like 1.49v for 5ghz. My lower vid hotter chip has 1.215v default Vid and runs 5 ghz @ 1.45v.