[HWiNFO32 Thread] - Questions, Reports, Suggestions, Discussions

I have much info, but there's no chance to read that all.. I need to know exactly what to search for..

 

^^As in referring to the duty cycles mentioned before. That's cool though, I really wouldn't know where to point you.


To try and explain what I had hoped to achieve back then.


This is a P8600 C2D Mobile CPU and as you can see the maximum multiplier is 9.5x with IDA. By using SLFM mode we can use a 19x multiplier. Now the idea was because the internal BCLK was half then the external BCLK could be doubled or more resulting in the internal clock being nearer its normal frequency of 266MHz. That coupled with up to 19x multi would give a possible (but unlikely) CPU range to >5GHz. Maybe because of the memory clock divider or something else internally, it did not work to plan. Also the higher multipliers >15x were less stable, whether this was due to PLL jitter idk. Long story short, it didn't work out. Anyway this is old tech, SB should bring some new challenges lol.

I notice with 9.5x 266 I get red bars but with 19x 133 they are yellow. Not sure if that was intentional, not that it really matters. Great software BTW.

 

Hm, I think there are several other factors that wouldn't allow such experiment - at least the Vcore, thermal dissipation, forcing SLFM always, etc. Btw, how have you changed the SLFM multiplier, using ThrottleStop?

The bars should be red only in case your clock is over the Max. Turbo (IDA) limit. If you attach a screenshot of such situation, I'll check if it's a bug.

 

Okay, must be while it's overclocked they turn red. I should have been more observant, sorry.

I was at the time interested to see how a lower powered mobile CPU vs its higher powered desktop counterpart would OC. If a 25W TDP CPU scales well then perhaps a 40% increase in Vcore to double the frequency would result in up to ~100W power draw at high load. This would not be for everyday use but just for interest to see how far the CPU could be pushed. All of the things you mentioned are currently inconsequential as the problem is the BCLK limit with FSB and RAM long before those other things become a big factor. I used my own software / driver for changing multipliers etc.

 

HWiNFO32 v3.62-1033 Beta available.
Changes:

• Enhanced support of ATI PM2 sensors.
• Added support of CHiL CH8266 PWM (for monitoring of GPU voltage, current, power).
• Added universal support of CSMI SAS (improved drive detection).
• Enhanced sensor monitoring on ASUS Rampage III GENE/Extreme.

 

HWiNFO32 v3.62-1039 Beta available.
Changes:

• Enhanced support of ATI PM2 sensors.
• Added support of CHiL CH8266 PWM (for monitoring of GPU voltage, current, power).
• Added universal support of CSMI SAS (improved drive detection).
• Enhanced sensor monitoring on ASUS Rampage III GENE/Extreme.
• Enhanced support of AMD Family 15h.
• Added support of Intel XTU v2.1.
• Added nVidia GeForce GTX 570.

 

Hi Mumak, is it possible to add some sort of support to change the fan speed for MSi notebooks?

I think a lot of the MSi gaming notebooks have the same EC.

Pretty sure this is the datasheet for the EC in my notebook:
http://wiki.laptop.org/images/a/ab/KB3700-ds-01.pdf

EDIT: Nope I was wrong, still looking for the correct datasheet.

This is my EC:
http://www.ene.com.tw/en/product_detail.asp?pid=366&id=2269

But I don't think I can find the datasheet.

 

In case it's really controlled by an EC, that might be tricky since it runs its own firmware and it's not compatible with any standard. Even having the datasheet won't help since that would require the firmware specification. Other machines I did had a standard fan speed controller, others required to disassemble the BIOS...
Is HWiNFO32 at least able to report the fan speed?

In order to investigate this further, I'd need some information:
- HWiNFO32 Report + Debug File including Sensors
- ACPI DSDT dump (you can use the RW-Everything Tool)
Even having all that information doesn't guarantee that I'll be able to do that.

 

No.

Ok I have attached a zip containing the info:

 

Unfortunately I can't see any traces of fan speed reporting or control there.
I see only two potential candidates for reporting temperature (CPU and System). The only thing I could do is to add them (that would require an EC dump using RW-Tool).

 

That would be nice, I have attached the dump:

 

Try this build:
www.hwinfo.com/beta/hw32_362_1040.zip
you should see a new sensor with 2 temperatures. Let me know how it works (you might try idle/load to check)... I'm not sure if there's anything else that might be reported...
Btw: This CPU doesn't seem to report DTS temperature which is probably due to the early ES1 stepping.

 

I don't see a new sensor, I have attached a report.

Should I try it with a CPU(T7300) that can report DTS temperature?

 

Sorry, please try this one:
www.hwinfo.com/beta/hw32_362_1041.zip
No need to use a different CPU..

 

Thankyou, that works.

I have attached a report.

I ran intelburntest for a while, the CPU sensor would just jump from 51C to 74C then to 83C. Guess it's not as good as a normal CPU sensor.

Much better than having nothing though, thanks again.

 

You're welcome
No idea how's that sensor implemented but it's quite common that temperature sensors are inaccurate. Even the DTS ones especially at room or lower temperatures...

 

The C2D idle temps can be adjusted to give better results providing they are not stuck or out by a stupid amount.

 

I experience some strange behaviour of HWiNFO32(1041).
If I start my notebook without an eSATA HDD attached, I get sensor information as shown in the left part of the first screenshot.
After attaching the eSATA drive the right picture is shown. This seems to be right for me.
If I start Win7 with the eSATA drive attached HWiNFO32 sensor information is shown as in thumbnail 2.
I have 2 Samsungs SSD in RAID0 and I think they have no temperature sensors.
What's wrong?

 

Yes, there seems something to be wrong. In the latest builds I have improved some methods for enumerating of RAID drives and reading of SMART values.
I'd need a Debug File from both situations to look more precisely.

 

here are the dbg files.
1. start hwinfo32 with no esata attached.
2. start hwinfo32 after attaching esata hdd.
3. start hwinfo32 after boot with attached esata.

 

Please try this build:
www.hwinfo.com/beta/hw32_362_1047.zip
and let me know how it worked.

 

Same result. When the eSATA is attached from the beginning the eSATA HDD and one Samsung SSD appear in the sensors window.
I have 2 Samsungs PM800 in RAID0, both of same capacity, but one is a FDE model capable of encryption.
Is this the cause?

 

No, it seems to be a problem of addressing RAID drives in HWiNFO32. I have to work on this a bit more.. It's a bit complicated when there are several drives using various interface connected. Will post a new update later.

 

Please try this build:
www.hwinfo.com/beta/hw32_362_1048.zip
and let me know how it works. In case there's still an issue, please attach a new Debug File.

 

1048 does work. Thank you very much!!
Another question:
The detail view of the 2 SSDs show

TRIM Command: Supported (Indeterminate Read After TRIM)

I thought with RAID TRIM is not supported? Am I wrong?

 

Thanks for the feedback.
The reported TRIM capability is per drive (and the drive itself supports it). I guess the availability of TRIM in RAID is a matter of software/RAID host. So in this case it's probably supported on drive, but not enabled on system (RAID) level.

 

HWiNFO32 v3.62-1049 Beta available.
Changes:

• Enhanced support of ATI PM2 sensors.
• Added support of CHiL CH8266 PWM (for monitoring of GPU voltage, current, power).
• Added universal support of CSMI SAS (improved drive detection).
• Enhanced sensor monitoring on ASUS Rampage III GENE/Extreme.
• Enhanced support of AMD Family 15h.
• Added support of Intel XTU v2.1.
• Added nVidia GeForce GTX 570.
• Enhanced sensor monitoring on Intel 6-series desktop boards.
• Added nVidia GeForce GT 540M.
• Enhanced support of Ivy Bridge-DT/MB.
• Added preliminary support of Ivy Bridge-EP/EX (Ivytown).
• Enhanced sensor monitoring on MSI GX620.

 

Is there any way to disable the screen that pops up when you launch the program that says "run" and "configure?" It's annoying to have to choose "run" every time. I mean, right after an install it makes sense. But...you don't want to configure it every time you run it...lol.

If there isn't a way, could you at least add an option to disable that screen?

Thanks.

Great work by the way.

 

There are 2 options available:
1. Either enable the "Minimalize Main Window on Startup" option, but this will start with the GUI not visible (in tray only)
2. Or open the HWiNFO32.INI file in a text editor (notepad) and in the "[Settings]" section set "ShowRegDialog=0".

 

NICE. +rep.

Thanks man. Keep up the great work!

 

HWiNFO32 v3.65 available.
Changes:

• Enhanced support of ATI PM2 sensors.
• Added support of CHiL CH8266 PWM (for monitoring of GPU voltage, current, power).
• Added universal support of CSMI SAS (improved drive detection).
• Enhanced sensor monitoring on ASUS Rampage III GENE/Extreme.
• Enhanced support of AMD Family 15h.
• Added support of Intel XTU v2.1.
• Added nVidia GeForce GTX 570.
• Enhanced sensor monitoring on Intel 6-series desktop boards.
• Added nVidia GeForce GT 540M.
• Enhanced support of Ivy Bridge-DT/MB.
• Added preliminary support of Ivy Bridge-EP/EX (Ivytown).
• Enhanced sensor monitoring on MSI GX620.
• Improved recognition of Arrandale ECC.
• Added support of SMSC SCH3112/4/6 LPC HW monitor.
• Fixed Benchmark comparison label issue.
• Added Intel Core 2nd generation logos.

 

Just updated...great job.
Had to uninstall the previous version and delete all traces...the icon wasn't showing up. But eventually worked just fine.

Got a question though...what's the persistent driver? I ticked that off...if you could shed some light on the advantages/disadvantages that would be great.

 

When you move the mouse cursor over the items in Configure screen it should give a quick help.
When Persistent driver option is enabled, the HWiNFO32 HW access driver is installed in system when run for the first time and remains there for subsequent runs. This should allow further launching without Administrator rights.
When disabled, then it's installed and removed from system each time it's run. This option is suitable for Portable packages.

 

is there anyway to get HWINFO32 to log only one sensor when i start logging?

 

All sensor data are logged, but you might try to disable the sensors/values you are not interested in (right-click and Disable).

 

Hi Martin, HNY2U.

I was curious if HWinfo32's reporting of IDA being supported is based on the actual fuse setting or something else?

 

Happy New Year too !
Reporting of IDA is based on the state of certain register(s) which I believe reflect the fuse state.

Btw, is there anyone here who's machine reports the CPU Power Consumption (in PCH Sensor)? I tried to improve the precision (not meaning that it wasn't precise already) and need to test this with a new Beta.

 

Do you mean bit 63 of 0xCE?

 

yep
<bla, bla... the message needs to be at least 10 chars long>

 

Where did you get that info?

Can you give me a link please?

 

Sorry moral, the only link I can give is to your quote as it's empirically derived and though I have suspected this bit was the IDA feature flag for some time I'm not sure any longer. It may be at best indirectly involved with IDA as for instance allowing certain c-states required for its use.

Martin, don't know if this is what you are looking for but I tried 365 on my desktop P55-UD6 and although it shows the "Intel PCH" sensor there is no data for it, ie blank.

 

If this bit/MSR is not described in the IA32 Software Dev.Manual, then it's under NDA only (so I should not talk about it anymore )

 

If it's under NDA and you can not discuss it that's cool but I'll give some information that while might not be very useful you might find interesting. You can just listen.


The P8400 does support IDA so why is it being shown as not supported.

Bit 63 is 0 and IDA does not work although IIRC the IDA FID:VID is still shown. So how did the P8400 end up in this state? Well it was run without a microcode update. Simply by running a microcode update later with software will return IDA operation and bit 63 will be set. This would suggest that bit is not tied to a fuse although it can still be a good indicator of IDA presence.

 

I listen
Well, there are some things inside the CPU that just very few ppl know how they work. And one of them is the uCode update (MCU) - the format of the update, what it can do/change, etc... This is one of the most kept secrets inside Intel and I guess you won't even get an answer to this
Some information about MCU is available under NDA, however this is just what you "need to know". Nothing more. From what I know so far is, that the MCU is very powerfull and can modify lots of things inside that can be modified (fix bugs, change behaviour). So if your dilemma is where IDA fuse/capability is really stored and how it can be modified, I can't answer this question. Maybe you could invite some Intel core team member for beer+vodka+wine+slivovica and when he is in good passion maybe he could tell you something But I doubt there's another way
EDIT: One more thing.. there's a single MCU per CPU Stepping/Platform, so it's not the case that the MCU you have is specially targeted for P8400.. Must be something general...

 

HWiNFO32 v3.66-1076 Beta available.
Changes:

• Fixed measuring of CPU clock on certain AMD CPUs under Windows 7.
• Improved displaying of CPU and GPU logo.
• Improved measurement precision of CPU Power consumption on Ibex Peak and Sandy Bridge platforms.
• Fixed recognition of Sandy Bridge K-series when used on platforms other than P67.

 

i found a big bug in this program if audio codec check is checked in settings there is a problem with Hdmi Audio out

 

Latest beta is listing my ati 5450 as a 5470 and showing the intel graphics logo instead of the ATI.

 

Please attach a Report File (and a screenshot of the problem) so I can check more precisely and fix it.

 

I've pmD u

 

Thanks. ATI does really weird things with their GPU branding.. One ID can have several different models. Can you please attach the Debug File (HWiNFO32.DBG) too? I need it to fix the GPU.