Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Solved Volterra VT1165 - Reading only VRM Phase2 Temp
#1
Though it's more of a nitpick than anything yet still merits reporting...

Issue:
When scanning the Volterra VT1165 IC, it will only report the second VRM Phase Temperature as 'GPU VRM Tempurature'. However, from what it seems it does infact give different readings for three phases using another utility such as Rivatuner using the VT1103.dll Plugin.

Notes:
I'm making an effort to reduce the probes between multiple apps to avoid issues from excessive probes [such as mis-read probe data due to a probe conflict between apps]. The only reason I'd taken note of this issue is because I've been using the Rivatuner plugin for Rainmeter to report the VRM Tempuratures and are the last item to convert my existing skin purely to HWiNFO monitoring through Stangowner's Rainmeter plugin.


Individual reports linked:
http://rwinds.no-ip.org/~archive/HWiNFO4...iNFO32.DBG
http://rwinds.no-ip.org/~archive/HWiNFO4...SCARAB.HTM

Comparison Screenshot linked:
http://rwinds.no-ip.org/~archive/HWiNFOx...ensors.png

Edit: Updated typo stating phase1 rather than phase2.
Reply
#2
Can you please attach also values of the other temperatures reported by the Rivatuner ?
Reply
#3
Meant to attach a screen shot comparison with the sensors divided rather than combined. Screenshot picture that is linked has been updated accordingly.
Reply
#4
Thanks. I'm working on it and will send you a test build in a while.
Are you able to determine which phase number temperature is currently reported by HWiNFO (1, 2 or 3) ?
Reply
#5
Just checked and ran Furmark to give a better separation of the sensors and it seems to most closely resemble Phase#2 actually rather than the originally reported Phase#1. Although #1/2 are normally within a degree of each other when the temps stabilize.
Reply
#6
Ok, thanks. Do you want to test HWiNFO32 or HWiNFO64 ?
Reply
#7
Please try this build of HWiNFO32: http://www.hwinfo.com/beta/hw32_407_1766.zip
and let me know how it works.

Please also note, that running multiple applications that communicate with the VT1165 concurrently might cause issues.
Reply
#8
Tested the build and it now shows Phase #1 and #2 accurately with phase 3 missing although 1/2 are mis-labeled as Phase 2 and Phase 3 respectively. Thanks for the effort thus far. Smile

Updated Dump:
http://rwinds.no-ip.org/~archive/HWiNFO4...iNFO32.DBG

Linked comparison shot:
http://rwinds.no-ip.org/~archive/HWiNFO1...ensors.png
Reply
#9
Thanks. Please try this one: http://www.hwinfo.com/beta/hw32_407_1767.zip
and report back Wink
Reply
#10
Just tested this build as well and bad news.. It works! *shakes head* ...wait.. that's GOOD news isn't it? Tongue

Anywho, seems things are in order here as the phases are accurately showing all 3 and in the proper assignments. Just verified through another torture test and only noticed a few inconsistent jumps but that's normal with probe collisions from synchronous reads due to two monitors running. Awesome work you've done done with HWiNFO.

Cheers!
Reply
#11
Haha, you scared me ! Smile
I'll do one more internal change to this, but I don't think it will require new testing (assuming I don't mess something Wink).
Reply
#12
(10-17-2012, 12:52 PM)Martin Wrote: Haha, you scared me ! Smile
I'll do one more internal change to this, but I don't think it will require new testing (assuming I don't mess something Wink).
Well, I can still do a test and related dump to give an extra check if needed to verify any regressions if it's of any help.
Reply
#13
OK, so here's the update (with dynamic readout of number of phases): http://www.hwinfo.com/beta/hw32_407_1768.zip
Please report back and enjoy Wink
Reply
#14
(10-17-2012, 01:05 PM)Martin Wrote: OK, so here's the update (with dynamic readout of number of phases): http://www.hwinfo.com/beta/hw32_407_1768.zip
Please report back and enjoy Wink

This build checks out also without any issues. No regressions and temps remain both spot on and properly assigned. I'll keep an eye on it to see if anything crops up at a later date. ^^

Nothing more annoying than optimizing something and it goes horribly wrong on the user-side.. been there before myself. Always a treat to have the opposite instead resulting in nothing other than good news. Thread has been marked as resolved accordingly.

Final Dump provided for confirmation of the results:
http://rwinds.no-ip.org/~archive/HWiNFO4...iNFO32.DBG
Reply
#15
Thanks for the feedback Wink
Reply
#16
Hmm.. after some more testing I noticed that the phases are being misreported over Low-Level I/O (NVAPI works fine) and instead matching temps with Phase #2. Came across this when adjusting settings to reduce time spent in the kernel (and thus cpu-use) between scans.

I apologize for not noticing this sooner during the initial testing of VRM Phase Support. x.x
Reply
#17
NVAPI is the preferred method of accessing GPU I2C, so please do not use the low-level method.
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)