| View previous topic :: View next topic |
| Author |
Message |
Johnscoty
Joined: 25 Aug 2012 Posts: 6
|
Posted: Sat Aug 25, 2012 11:31 am Post subject: What did I get? 1045t marked as 1040t and higher TDP |
|
|
I bought a 1045t off ebay (only the cpu, no box no fan). It works fine even under stress, benchmarks are fine, and the id numbers on the cpu are correct. But cpu-z calls it a 1040t (CPU -> Processor -> Name) and the tdp is over a hundred when all 1045t are 95W. Here is a screenshot:
I tried googling for 1040t and searching here for 1040t but got no answer. All cpu-z screenshots of a 1045t Ive found have 1045t under both Name and Specification and lower tdp.
My motherboard has cpu unlocking set off, it is possible that this is an unlocked cpu transformed into a 1045t? Everything Ive read says that to unlock cores in a cpu you need the bios and I dont think it is posible to do it directly on the cpu, is it? Another option Im thinking is that it could be a cpu from early AMD testings and thats why it has a higher tdp and the 1040t name? I really have no idea.
Both the cpu benchmarks and temperature is fine, I have not managed to get it over 45ºC even when its summer hot here and it has a cheap 20euros cooler (no oc). I can not test power consumption, but voltage seems fine. Still, I would like to know what did I get, why the higher tdp and the 1040t name (which I can not find any valid reference on google or anywhere). Thanks. |
|
| Back to top |
|
 |
gshv

Joined: 01 Feb 2003 Posts: 7898 Location: Fairfax, VA USA
|
Posted: Sun Aug 26, 2012 3:57 pm Post subject: |
|
|
There is no 1040T CPU. The "Specification" field shows CPU name generated by BIOS, so the BIOS correctly recognizes your CPU. The "Name" field is generated by CPU-z, so in my opinion it's a bug in CPU-z identification code. I don't know how CPU-z calculates or determines TDP, so it's difficult for me to comment on it. The CPU-z may pull this value from power-state configuration data from MSR registers. The fact that CPU-z calls it "Max TDP" tells me that they probably scan all power-states, find the largest TDP and report it. Well, Phenom II processors do have higher 10% - 15% higher TDP when they run in a single-plane socket (socket AM2), so 107W fits right in the middle of this range. So, your processor looks fine, and you shouldn't worry about those 1040T and 107W values.
Gennadiy |
|
| Back to top |
|
 |
Johnscoty
Joined: 25 Aug 2012 Posts: 6
|
Posted: Sun Aug 26, 2012 5:16 pm Post subject: |
|
|
Thanks for the answer.
| gshv wrote: | | The "Name" field is generated by CPU-z, so in my opinion it's a bug in CPU-z identification code. |
The thing is other programs give the same result as cpu-z. Name has 1040T while Specification has 1045T. Las one I checked is http://www.piriform.com/speccy for example, same results as CPU-z.
| gshy wrote: | | I don't know how CPU-z calculates or determines TDP, so it's difficult for me to comment on it. The CPU-z may pull this value from power-state configuration data from MSR registers. The fact that CPU-z calls it "Max TDP" tells me that they probably scan all power-states, find the largest TDP and report it. Well, Phenom II processors do have higher 10% - 15% higher TDP when they run in a single-plane socket (socket AM2), so 107W fits right in the middle of this range. So, your processor looks fine, and you shouldn't worry about those 1040T and 107W values |
Could very well be, but what keeps me unesay is that all the screenshots Ive found on google of "cpu-z 1045T" have lower TDP: http://www.google.es/search?q=1045t+cpu-z&sugexp=chrome,mod%3D4&um=1&ie=UTF-8&hl=es&tbm=isch&source=og&sa=N&tab=wi&ei=QZ86UNy4NsjRhAf6k4CADA&biw=1680&bih=949&sei=RJ86UILcFs6WhQeJ64DgAg (the only image that doesnt is mine, that has already been detected by google). |
|
| Back to top |
|
 |
gshv

Joined: 01 Feb 2003 Posts: 7898 Location: Fairfax, VA USA
|
Posted: Sun Aug 26, 2012 6:12 pm Post subject: |
|
|
What is a make and model of your motherboard?
Gennadiy |
|
| Back to top |
|
 |
Johnscoty
Joined: 25 Aug 2012 Posts: 6
|
Posted: Mon Aug 27, 2012 7:18 am Post subject: |
|
|
| gshv wrote: | | What is a make and model of your motherboard? |
The motherboard is a Gigabyte GA-890FXA-UD5 Rev 2.1 (just in case, it has lastest bios, version F6). |
|
| Back to top |
|
 |
gshv

Joined: 01 Feb 2003 Posts: 7898 Location: Fairfax, VA USA
|
Posted: Tue Aug 28, 2012 1:36 am Post subject: |
|
|
Oh, so it's a socket AM3 board... I have no idea why it reports wrong TDP. It could be using core voltage and current to calculate this value. Try to slightly reduce core voltage in BIOS and see if that changes anything.
Gennadiy |
|
| Back to top |
|
 |
Johnscoty
Joined: 25 Aug 2012 Posts: 6
|
Posted: Tue Aug 28, 2012 6:26 am Post subject: |
|
|
| gshv wrote: | Oh, so it's a socket AM3 board... I have no idea why it reports wrong TDP. It could be using core voltage and current to calculate this value. Try to slightly reduce core voltage in BIOS and see if that changes anything.
Gennadiy |
I can not mess with the BIOS now, I will try later. But it is actually consuming more than it should. According to this document: http://globalsp.ts.fujitsu.com/dmsp/Publications/public/wp-energy-ESPRIMO-E5645-E85.pdf (page 3), the 1045T should consume between 48W (idle) and 94W (maximum).
I installed CPUID Hardware monitor and it tells you how much the CPU is consuming:
Idle matches (48.3 W) but maximum goes up to 109.2W which goes in line with the higher TDP reported and its way higher than the theoretical maximum of 94W.
I really dont have much experience at this level with CPU, does anyone know if this is usual or there is something going on? I will try lowering the voltage later on, since the minimum voltage is a bit higher than the reported value (http://www.cpu-world.com/CPUs/K10/AMD-Phenom%20II%20X6%201045T%20-%20HDT45TWFK6DGR.html, says lower VCORE is 1.075V and in my case minimum is 1.150V, max also theoretically 1.375V- 1.425 on turbo mode, mine 1.410V) |
|
| Back to top |
|
 |
Johnscoty
Joined: 25 Aug 2012 Posts: 6
|
Posted: Tue Aug 28, 2012 4:53 pm Post subject: |
|
|
Quick update: I undervolted the cpu (and tested undervolting the nbcpu as well). I have tried -0.1V and it seems to be working fine. I have not tested enough time (I will do later), but the early tests shows that it does not change much the consumption idle (4W less) but under load it saves 20W. The funny part is that performance seems to be unaffected. Ive run benchmarks and they give similar results.
Im still wondering why I see 1040T but the consumption part at least is solved. Thanks. |
|
| Back to top |
|
 |
gshv

Joined: 01 Feb 2003 Posts: 7898 Location: Fairfax, VA USA
|
Posted: Tue Aug 28, 2012 7:41 pm Post subject: |
|
|
| Johnscoty wrote: | | Im still wondering why I see 1040T |
It's more of a question to CPU-z developers. Have you tried CWID.exe from this site?
Gennadiy |
|
| Back to top |
|
 |
Johnscoty
Joined: 25 Aug 2012 Posts: 6
|
Posted: Wed Aug 29, 2012 1:01 am Post subject: |
|
|
| gshv wrote: | | Johnscoty wrote: | | Im still wondering why I see 1040T |
It's more of a question to CPU-z developers. Have you tried CWID.exe from this site?
Gennadiy |
Just did, it says 1045T. |
|
| Back to top |
|
 |
|