Overclock.net banner
1 - 12 of 12 Posts

·
Registered
Joined
·
50 Posts
Discussion Starter · #1 ·
Rectangle Slope Font Line Parallel


HI

the spec,
PSU : Cooler master 550W
Montherboard : Gigabyte ud3r p45
Gpu old 980 Ti
Cooler GALAHAD AIO 360
8500 ddr2

Test Prime 95 small FTT for 30 minute, 1.3V 1.4V and 1.45-1.475V bios


the problem only one e8500 yellow one, is the only one that can do 4.6 ghz, under 1.45V. Some cpu can do 4.6 ghz at over 1.5V

the green cpus have better silicon quality than the 4.6 ghz E8500 and should be able to do 4.6 ghz. the last one can do 4.5 ghz at 1.3V(1.26reel) but none of the golden green cpu can do 4.6 ghz, no matter the voltage.

the green cpu run hotter than the rest and my cooler cannot keep up, after 4.5 ghz the cpu can go above 80C. i believe that why i cannot go to 4.6 ghz with the green cpu. but i don't understand why a 360 AIO cannot keep up. i feel like i hit a wall.

am i missing someting ?
Do i just need a better cooler and what cooler can be better than a 360 AIO ?
why the golden cpu run so hot and the crap cpu can have 1.55V no problem ?
 

·
Banned
Joined
·
1,185 Posts
View attachment 2543883

HI

the spec,
PSU : Cooler master 550W
Montherboard : Gigabyte ud3r p45
Gpu old 980 Ti
Cooler GALAHAD AIO 360
8500 ddr2

Test Prime 95 small FTT for 30 minute, 1.3V 1.4V and 1.45-1.475V bios


the problem only one e8500 yellow one, is the only one that can do 4.6 ghz, under 1.45V. Some cpu can do 4.6 ghz at over 1.5V

the green cpus have better silicon quality than the 4.6 ghz E8500 and should be able to do 4.6 ghz. the last one can do 4.5 ghz at 1.3V(1.26reel) but none of the golden green cpu can do 4.6 ghz, no matter the voltage.

the green cpu run hotter than the rest and my cooler cannot keep up, after 4.5 ghz the cpu can go above 80C. i believe that why i cannot go to 4.6 ghz with the green cpu. but i don't understand why a 360 AIO cannot keep up. i feel like i hit a wall.

am i missing someting ?
Do i just need a better cooler and what cooler can be better than a 360 AIO ?
why the golden cpu run so hot and the crap cpu can have 1.55V no problem ?
Because of the transistor current leakage, inherent of how silicon doped chips work.

When speaking about photo-lithography nodes:
-A low leakage chip would require less voltage to run, but output much more heat when running, at the same clocks.
-A high leakage chip would require more voltage to run, but output much less heat when running, at the same clocks.

That's why extreme overclockers usually looked for good but high leakage cpu bin.
Because one can push more voltage, which mean higher clocks, when cooled properly.
Instead, people that wanted to run a simple air cooler, would prefer a low leakage cpu bin.
Because, if not pushed to the max, it will run cooler at a low voltage, with decent clocks.

So yes, in any case, brute forcing the cooling solution is typically the best and easiest way, to push the clocks up.
 

·
Registered
Joined
·
354 Posts
Why do you need to overclock to 4.6GHz? Could be silicon limitation or maybe liquid nitrogen could help. The CPU is so obsolete that it's fit only for modern web-browsing and performance gain over 4.5GHz would be negligible.
 

·
Registered
Joined
·
50 Posts
Discussion Starter · #4 ·
Because of the transistor current leakage, inherent of how silicon doped chips work.

When speaking about photo-lithography nodes:
-A low leakage chip would require less voltage to run, but output much more heat when running, at the same clocks.
-A high leakage chip would require more voltage to run, but output much less heat when running, at the same clocks.

That's why extreme overclockers usually looked for good but high leakage cpu bin.
Because one can push more voltage, which mean higher clocks, when cooled properly.
Instead, people that wanted to run a simple air cooler, would prefer a low leakage cpu bin.
Because, if not pushed to the max, it will run cooler at a low voltage, with decent clocks.

So yes, in any case, brute forcing the cooling solution is typically the best and easiest way, to push the clocks up.
thx for your answer

the 360 AIO is not enough for the high leakage chip E8600, do i need to go custom loop ?, i will be my first custom if i do that. i think the 1.3V 4.5 ghz e8600 have potential
 

·
Registered
Joined
·
1,914 Posts
360 AiO is plenty enough for a dual core chip and if it's not, then it's most likely the CPU not giving away it's heat quickly enough.
 

·
Banned
Joined
·
1,185 Posts
not extreme overcloking, but this cpu is the best of the 30 cpu that i tried to oc. i want to push at is limit, maybe reach 4.7 ghz or 4.8 ghz. this is a projet that i do for fun
If used just for benching, you can try to submerge the AIO radiator into cold water, or taking out the rig at night.
The point being, finding an easy way to drop the AIO T° without messing with the system too much.
You can find some videos, where users dropped the AIO radiator into an ice water bath, to be able to reach lower T° when benching.

If you want to go a bit further, you should build a small custom loop, with a 775 compatible waterblock.
You can use a cheap irrigation pump and reuse the radiators of some old decommissioned AIO.

The next step would be to lap the cpu and the waterblock to be flat, to get the best surface match possible, hence the best thermal transfer.
And once again, you can drop the T° using an ice bath or a dry ice bath.

Going down below ambient T° have some drawbacks, like the condensation that could build up on the computer parts, resulting in short-circuits.
Being sure to use a proper glycol based coolant, to avoid water freezing in the loop.
And the fact that some bios and/or board modification are needed to be able to run below 0°.
 

·
Registered
Joined
·
50 Posts
Discussion Starter · #9 · (Edited)
If used just for benching, you can try to submerge the AIO radiator into cold water, or taking out the rig at night.
The point being, finding an easy way to drop the AIO T° without messing with the system too much.
You can find some videos, where users dropped the AIO radiator into an ice water bath, to be able to reach lower T° when benching.

If you want to go a bit further, you should build a small custom loop, with a 775 compatible waterblock.
You can use a cheap irrigation pump and reuse the radiators of some old decommissioned AIO.

The next step would be to lap the cpu and the waterblock to be flat, to get the best surface match possible, hence the best thermal transfer.
And once again, you can drop the T° using an ice bath or a dry ice bath.

Going down below ambient T° have some drawbacks, like the condensation that could build up on the computer parts, resulting in short-circuits.
Being sure to use a proper glycol based coolant, to avoid water freezing in the loop.
And the fact that some bios and/or board modification are needed to be able to run below 0°.
that seem like a pretty good idea.



i will try the ice bath but first i will test the rest of my cpu in normal, and after worthy (green one) ,i will test them on the ice bath.


the majority of E8XXX are more silicon limited than temperature
Another E8500 reach 4.6 ghz at 1.475V)(1.440V cpu-z)


For now 36 cpu
100% can do 4.1 ghz
97% can do 4.2 ghz (an E8600 failed)
80% can do 4.3 ghz
61% can do 4.4 ghz
36% can do 4.5 ghz
5.5% can do 4.6 ghz (only 2, 4 above 1.5V)
 

·
Banned
Joined
·
1,185 Posts
that seem like a pretty good idea.



i will try the ice bath but first i will test the rest of my cpu in normal, and after worthy (green one) ,i will test them on the ice bath.


the majority of E8XXX are more silicon limited than temperature
Another E8500 reach 4.6 ghz at 1.475V)(1.440V cpu-z)


For now 36 cpu
100% can do 4.1 ghz
97% can do 4.2 ghz (an E8600 failed)
80% can do 4.3 ghz
61% can do 4.4 ghz
36% can do 4.5 ghz
5.5% can do 4.6 ghz (only 2, 4 above 1.5V)
Have no other words that, awesome job! (y)

Well, i admit there is a sharp dip between 4.5Ghz and 4.6Ghz.
By the way, you should check the Luumi channel, he has a nice playlist of e8600/e8500 binning and overclocking.
 

·
Registered
Joined
·
50 Posts
Discussion Starter · #11 · (Edited)
Have no other words that, awesome job! (y)

Well, i admit there is a sharp dip between 4.5Ghz and 4.6Ghz.
By the way, you should check the Luumi channel, he has a nice playlist of e8600/e8500 binning and overclocking.
thx for your help

i tried the ice water bath, work well 2 of the golden cpu managed 4.6 ghz, it was a good idea. the setup was not optimal but i lost 10C with is very good.


Rectangle Slope Font Line Parallel


my best cpu the last E8600 was to hot even with the ice bath for 4.6 ghz. i will try again,

4.7 ghz is way harder than 4.6 ghz and the cooling become an issue again.I am also not sure how much the gigabyte ud3r p45 can take
 

·
Banned
Joined
·
1,185 Posts
thx for your help

i tried the ice water bath, work well 2 of the golden cpu managed 4.6 ghz, it was a good idea. the setup was not optimal but i lost 10C with is very good.


View attachment 2548412

my best cpu the last E8600 was to hot even with the ice bath for 4.6 ghz. i will try again,

4.7 ghz is way harder than 4.6 ghz and the cooling become an issue again.I am also not sure how much the gigabyte ud3r p45 can take
Awesome stuff as usual. (y)
I'm very happy you managed to break your clock limit.

Well, i guess the next step is lapping or deliding the cpu, alongside lapping the AIO block.
You should be able to shave off, a couple of ° more, when lapping both surfaces.

Deliding would be a bit more difficult, because this old cpu is soldered.
Tho, i expect the gallium solder being quite old and more fragile, at this point in time.
Going direct die would be the best scenario, especially if you keep going on, with the dry ice, ice water bath.

I also agree, the motherboard will soon reach its limit i suppose, the ud3r p45 runs a 6 phases cpu vrm.
The more you push the voltage, the more the vrm heats up.
Still, if actively and properly cooled, you should still be able to push a bit more.
I guess than a high-end board would be more welcomed, at this point.
 
1 - 12 of 12 Posts
Top