GPU-Z 0.2.7 reading

[darkfrank]

Can anyone tell me wt is the MEMIO reading for.....?
Cuz i got some trouble after installed my AC s1 on 4870,with all ram heatsink and mosfet heatsink,i got this

 

[alexp999]

Pic not workie.

 

[darkfrank]

ok it works now.

 

[alexp999]

Wow thats a lot of temps, i thought the 48xx series only had 3 temps. All on GPU core. Dont like the fact that your fan is on 1rpm. is it spinning okay??

 

[newtekie1]

It is one of the three on die temp sensors in the RV770 core. I don't know if it is normal for it to be that high, maybe someone with 4800 can tell you what their readings are.

 

[darkfrank]

Wow thats a lot of temps, i thought the 48xx series only had 3 temps. All on GPU core. Dont like the fact that your fan is on 1rpm. is it spinning okay??

I have installed my AC S1 with a 12CM fan,so it does not show the fan speed.

 

[darkfrank]

It is one of the three on die temp sensors in the RV770 core. I don't know if it is normal for it to be that high, maybe someone with 4800 can tell you what their readings are.

I got a reboot if the gpu was under full load...so probably the MEMIO temp is too high,just wanna know which part it was

 

[alexp999]

They are all on the GPu die afaik. you may need to reseat the cooler. What thermal paste did you use?

 

[FR@NK]

You need to make sure the cooler base has even pressure on the GPU die and is making good contact. Also make sure the TIM is evenly spread over the die. The MEMIO is normally the hottest of the 3 core temps. It should be 5-15C hotter then the other two.

 

[nafets]

I agree with alex and Fr@nk. You need to re-seat/re-install your S1. The temperature variation between the three on-core GPU temps is much too great, signifying poor contact between the S1 base and the GPU core...

 

[W1zzard]

yep the problem is that your cooler is making uneven contact with the die. since the memio sensor is not located in the middle of the die it wont get cooled as well if the baseplate isnt sitting right on the core

 

[acrowley23]

Hi

Nice work! 0.27 works perfect on my 4870Card

can someboy explain the new values ?
(not my card on this pic)

[/QUOTE]

Is it ?:
1: GPU Temp
2. Memory temp
3. Shadercore ( what is shadercore?)

VDDC=VoltageRegulators /Vitec/ Pulse regulators ?

Thank you

 

[LuxZg]

Hi

Nice work! 0.27 works perfect on my 4870Card

can someboy explain the new values ?

Is it ?:
1: GPU Temp
2. Memory temp
3. Shadercore ( what is shadercore?)

VDDC=VoltageRegulators /Vitec/ Pulse regulators ?

Thank you

Well, my guess/logic is:
GPU DISPIO - GPU (meaning GPU chip itself) DISPlay Input/Output; this would be part of GPU core that is responsible for communication with components that lead to your monitor (display)
GPU MEMIO - GPU MEMory Input/Output; meaning this is the temperature of the memory controler that is a part of GPU chip
GPU SHADERCORE - well, not much to explain. Shader core is part of GPU chip (core) that contains shader processing units, so this is the temperature of processors of your graphics card

VDCC Slave 1/2/3 - I have no exact idea, these are obviously VRM regulation temperatures, just not sure what slave 1/2/3 means. Are these temperatures for three chips consisting voltage regulation of the card?

Now what I'd really like is a scheme of GPU and VRM, showing which area sensors are placed in. Wizz, you seem to have that info It would help a lot in situations like darkfrank has, where only one temperature reading is "off scale" Thanks

 

[W1zzard]

luxzg is correct about the 3 on-die sensors

the vddc slave temps are the temperatures of the 3 pwm phase controllers on the card. marked on the picture below, hard to see. to the right of the big voltage controller ic

 

[Lionheart]

His mine, what do you thinks of the temps, I tried overclocking it ran well but I dont trust this motherboard, it only has PCI E not 2.0

 

[LuxZg]

@Chaos - seems fine to me..a bit weird that shadercore is 5C higher (usually first and third reading were always pretty much the same, only second being a bit higher, at least on 4870's if I remmember right). But since it's pretty cool 45-50C on all three in idle, I'd say it's all fine.

@Wizz..

luxzg is correct about the 3 on-die sensors

the vddc slave temps are the temperatures of the 3 pwm phase controllers on the card. marked on the picture below, hard to see. to the right of the big voltage controller ic
(img cut out)

Ummm..marked on the picture, and hard to see? I can't see sh**
Oh, right, I saw it now when I've saved it and opened in paint with 2x zoom You should have done it in green, or any other non-PCB-red color But thanks for confirming

Btw, do you know anything about which is slave 1 and which is slave 3? I guess slave2 should be in the middle..
And I wonder.. will this card show 4 such temperatures, as small (but noticable) differences seen in last picture confirm it has a bit longer VITEC voltage controller chip (populates that empty space with free connections on PCB), and additional components where 4th "vddc slave" chip is missing on reference cards

 

[nafets]

Spent an hour goofing around with the VRM temps;

-----------------------

His mine, what do you thinks of the temps, I tried overclocking it ran well but I dont trust this motherboard, it only has PCI E not 2.0

Temps look fine for your HD4850. Although I don't know how you can stand running the fan at that speed. I'd be ready to kill someone with that droning.

The functionality and performance of your HD4850 won't be affected by the PCI-E revision (1.x/2.0) of the slot on your motherboard.

 

[W1zzard]

And I wonder.. will this card show 4 such temperatures, as small (but noticable) differences seen in last picture confirm it has a bit longer VITEC voltage controller chip (populates that empty space with free connections on PCB), and additional components where 4th "vddc slave" chip is missing on reference cards

yes, unless powercolor messed up the voltage regulator design somehow

 

[acrowley23]

@wizzard

is its possible to read out the Temp from VITEC ?

And what is the function from the pwm phase controllers ?
I thought these 3 Chips are the Voltage Regulator / GPU VRM.
In EVEREST its called GPU VRM

And can you add a monitoring/logging -save to file Function please ?

 

[W1zzard]

possible: no, there is no point, it won't get hot
the voltage controller ic has no current going through it, it just controls the pwm phases.
these three chips are part of the gpu vrm, yes

i googled some random voltage controller:

marked red are the six mosfets (2 each are in the 3 phase controller ics)

 

[acrowley23]

Ok, THX

As i say....are planning some logging Function for the GPU-z monitoring ? Would be perfect

 

[W1zzard]

yes, i can add logging. give me your wish list what features you want

 

[nafets]

The large Vitec chip and the two smaller Pulse chips are effectively cooled without heatsinks, with just some low/moderate airflow from the VGA cooler fan or a nearby case fan. You shouldn't have to worry about them too much...

 

[acrowley23]

yes, i can add logging. give me your wish list what features you want

nice

my wish is nothing special :
- logging of all Sensors/ option to select custom Sensors
- save to file
- monitoring Graph (somethig like Rivatuners monitoring Window)
- GPU Z logitech G15 Applet with monitoring / vista Sidebar

 

[LuxZg]

I second those requests But logging should be fine at first

And I have one more:
- give us the option to resize GPU-Z window ourselves.. I want to see all sensors at once, without scroll.. Especialy since I can see 13 out of 14 anyway, so I have to scroll just for that one last sensor

@nafets - thanks for the markings of 1/2/3, and just to make it complete.. Is this what you ment about heating?