Tom's Fan Failure Test

[21stHermit]

Tom's Hardward did a four CPU comparison whereby they removed the fan wire from the MB and then ran a 10 set benchmark. The only CPU which ran all 10 was the E2160.

The other CPU's: AMD A64 X2 6000+, Athlon X2 BE-2350, Intel E6850.

"the Pentium Dual Core E2160, which is a low-cost processor rated at a TDP of 65 W, managed to complete all ten benchmarks with the CPU fan disconnected!"

The full multi-page article for your detailed review.

Hermit

 

[AmberClad]

Hmm, your topic summary is a little inaccurate. None of the CPUs in that test "died". Three of the four simply shut down, like they're designed to. No catastrophic failures like during that test THG did a few years ago.

 

[nyker96]

that's pretty impressive for E2160! now let's see if a fanless setup works for this guy, maybe a TR 120 UX can cool this guy fanlessly.

 

[ArchAngel777]

Originally posted by: AmberClad
Hmm, your topic summary is a little inaccurate. None of the CPUs in that test "died". Three of the four simply shut down, like they're designed to. No catastrophic failures like during that test THG did a few years ago.

Agreed, terribly misleading topic summary.

 

[gOJDO]

If they undervolted and underclocked the rest of the CPU's to a E2160's frequency I am pretty sure that all would accomplished the same tasks.

 

[Viper GTS]

FWIW I have a BE-2300 system running fanless on a GeminII.

The only fan in the system is the 120mm in the Seasonic S12II.

It runs quite toasty (I need to loosen the mount on the GeminII as I may have over-tightened it) but it's orthos stable for hours on end. Needless to say it is one quiet system.

Viper GTS

 

[Accord99]

Not suprising since the Pentium E2160 is the 2nd lowest power-consuming non-mobile dual-core CPU on the market today:

http://www.digit-life.com/arti...-e2140-2160-page3.html

 

[Arkaign]

How does the 2140 compare?

 

[Accord99]

digit-life has the 2140 consuming 1W less under load.

 

[Arkaign]

Originally posted by: Accord99
digit-life has the 2140 consuming 1W less under load.

So it's safe to assume that it would behave the same under this test? What of the 2180?

 

[21stHermit]

Originally posted by: Accord99
Not suprising since the Pentium E2160 is the 2nd lowest power-consuming non-mobile dual-core CPU on the market today:

http://www.digit-life.com/arti...-e2140-2160-page3.html

Thanks for that link. :beer:

Good info.

 

[Accord99]

Originally posted by: Arkaign

Originally posted by: Accord99
digit-life has the 2140 consuming 1W less under load.

So it's safe to assume that it would behave the same under this test? What of the 2180?

Yep, the E2140 would pass the test too. I haven't seen any reviews of the E2180 though it looks like it has the M0 stepping. A Belgium review of the M0 E4500 suggests that this stepping slightly reduces power vs the L2 stepping:

http://www.matbe.com/articles/...m-et-celeron/page5.php

 

[PetNorth]

Originally posted by: gOJDO
If they undervolted and underclocked the rest of the CPU's to a E2160's frequency I am pretty sure that all would accomplished the same tasks.

yeah.

Toms hardware tests, as foolish as usual.

 

[Arkaign]

Originally posted by: PetNorth

Originally posted by: gOJDO
If they undervolted and underclocked the rest of the CPU's to a E2160's frequency I am pretty sure that all would accomplished the same tasks.

yeah.

Toms hardware tests, as foolish as usual.

Of course, there's a difference between a processor running at it's original speed, and one crippled by underclocking. I wouldn't call it foolish at all, but a good test for people who want to try passive cooling for HTPC or silent solutions.

 

[PetNorth]

Originally posted by: Arkaign

Originally posted by: PetNorth

Originally posted by: gOJDO
If they undervolted and underclocked the rest of the CPU's to a E2160's frequency I am pretty sure that all would accomplished the same tasks.

yeah.

Toms hardware tests, as foolish as usual.

Of course, there's a difference between a processor running at it's original speed, and one crippled by underclocking. I wouldn't call it foolish at all, but a good test for people who want to try passive cooling for HTPC or silent solutions.

not at all. We don't know real voltages (some mobos overvolt cpu with voltage set to "auto", others undervolt), we don't know temperatures, in fact we know basically nothing about the accuracy of this "test".

 

[Keysplayr]

Originally posted by: gOJDO
If they undervolted and underclocked the rest of the CPU's to a E2160's frequency I am pretty sure that all would accomplished the same tasks.

What voltage and speed is the BE-2300 running?

 

[secretanchitman]

Originally posted by: keysplayr2003

Originally posted by: gOJDO
If they undervolted and underclocked the rest of the CPU's to a E2160's frequency I am pretty sure that all would accomplished the same tasks.

What voltage and speed is the BE-2300 running?

i was gonna ask the same thing.

 

[Keysplayr]

BE-2300 is 1.9GHz
Core Voltage:
1.15- 1.20 V

E2160 is 1.8GHz
Core voltage:
1.225-1.325 V

So this statement: "If they undervolted and underclocked the rest of the CPU's to a E2160's frequency I am pretty sure that all would accomplished the same tasks." Is out the window.
Are you still pretty sure?

So, you can underclock the BE-2300 100MHz, but you'd have to increase the voltage to 1.225 or higher to equal the E2160. Or did you just want to underclock?

 

[BladeVenom]

If running a processor without a fan is important to you then VIA rules.

 

[Keysplayr]

For most folks with HTPC's, quieter is better. Others, not so much.

 

[AmberClad]

Originally posted by: keysplayr2003
BE-2300 is 1.9GHz
Core Voltage:
1.15- 1.20 V

E2160 is 1.8GHz
Core voltage:
1.225-1.325 V

I believe the THG test used the BE-2350 (2.1 GHz), not the BE-2300.

Even stranger, if you look at the new AT article posted this morning, the test BE-2350-based system actually pulls less watts than the E2160 system, both at idle and at load:

BE-2350
Idle: 49.6W
Load: 75.3W

E2160
Idle: 60.1W
Load: 81.4W

Almost lead me to believe that there was something wonky going on with Tom's testing methodology, since his results are counterintuitive based on Anand's findings. But the last page of Anand's article has OCing results:

E2160: 67.70% OC, from 1.8GHz to 3.02Ghz
BE-2350: 14.70% OC, from 2.1GHz to 2.41Ghz

The BE-2350 seems like an underachiever, I guess. Not sure what else to make of it .

 

[Accord99]

The 7050 based AMD motherboard uses less power than the G33 based Intel motherboard.

 

[Zap]

Originally posted by: BladeVenom
If running a processor without a fan is important to you then VIA rules.

Not for long. This thread mentions the upcoming Intel D201GLY2 board. The original D201GLY board sells for $69 at Frys.com. That's around $77 shipped to me (YMMV) for a mini ITX board with a mobile ULV Celeron. The new one is reportedly to be of similar cost as the old one, plus adds SATA ports and is passively cooled. I think that a Celeron 1.2GHz (basically a Core Solo with less cache and no SpeedStep).

Too bad network is not GBe.

 

[PetNorth]

Originally posted by: AmberClad

Originally posted by: keysplayr2003
BE-2300 is 1.9GHz
Core Voltage:
1.15- 1.20 V

E2160 is 1.8GHz
Core voltage:
1.225-1.325 V

I believe the THG test used the BE-2350 (2.1 GHz), not the BE-2300.

Even stranger, if you look at the new AT article posted this morning, the test BE-2350-based system actually pulls less watts than the E2160 system, both at idle and at load:

BE-2350
Idle: 49.6W
Load: 75.3W

E2160
Idle: 60.1W
Load: 81.4W

Almost lead me to believe that there was something wonky going on with Tom's testing methodology, since his results are counterintuitive based on Anand's findings. But the last page of Anand's article has OCing results:

E2160: 67.70% OC, from 1.8GHz to 3.02Ghz
BE-2350: 14.70% OC, from 2.1GHz to 2.41Ghz

The BE-2350 seems like an underachiever, I guess. Not sure what else to make of it .

thats the reason I said we know nothing about accuracy toms test. And I said it, because I owned an Asus M2N deluxe and It overvolted processor like hell (If I set voltage to auto, and cpu vcore default is 1.35, it overvolted it to 1.45v, figure it. If I set manually to for example 1.30, it overvolted to 1.40, etc). My guess is that Asus Crosshair makes the same...

Thats probably the main reason why in Anand test, BE- 2350 system consumes less power, because 690G boards (I own two of them) usually set cpu voltage correctly.

 

[aigomorla]

Originally posted by: nyker96
that's pretty impressive for E2160! now let's see if a fanless setup works for this guy, maybe a TR 120 UX can cool this guy fanlessly.

uhhh i have a E6300 cooled by an Ultra120EX passively. >.<


It doesnt do tons of encoding. Its a media center that basically plays movies, and streams off my NAS. But i havent had any problems with it yet.