1st Play With A 95 Watt 1055T

Balb0wa · 10 Aug 2010 at 09:52

1st settings ive tried, not bad for a £146 cpu.

HeX · 10 Aug 2010 at 09:54

Do 20 loops of LinX @ problem size 20,000.

Impressive if it manages that

Martini1991 · 10 Aug 2010 at 10:10

You're looking at the wrong temps lol.

gurusan · 10 Aug 2010 at 10:12

Very nice. A hex-core within that thermal envelope is impressive, and looks to be a good clocker too.

Balb0wa · 10 Aug 2010 at 10:18

Martini1991 said:
You're looking at the wrong temps lol.

i was looking at the vid, you can look at what you want

ive had a 1055T since they came out, i know to use hardware monitor for better temps.

Martini1991 · 10 Aug 2010 at 11:18

Balb0wa said:
i was looking at the vid, you can look at what you want

ive had a 1055T since they came out, i know to use hardware monitor for better temps.

Makes sense, just there was nothing else there to show your temperatures.

crisbduck · 10 Aug 2010 at 12:05

=( want mine soon!

gareth170 · 10 Aug 2010 at 13:56

Martini1991 said:
just there was nothing else there to show your temperatures.

core temp at 46c so i guess real cpu temp was at ~57c

Balb0wa · 10 Aug 2010 at 14:02

gurusan · 10 Aug 2010 at 15:21

So how about some stability testing?

HeX · 10 Aug 2010 at 15:25

Indeed, I wanna see 20 loops of LinX at 2048MB at least

gareth170 · 10 Aug 2010 at 15:28

prime test is better tbh.

HeX · 10 Aug 2010 at 15:30

gareth170 said:
prime test is better tbh.

Not IMHO, i've had prime run for 8hr's solid and LinX/IBT fail in 5mins.

LinX/IBT also heats your CPU up more than Prime.

I usually do 20 loops of LinX/IBT @ 2048MB (or problem size 20,000), if it passes that (takes about an hour) I give it the 8hr Prime Blend treatment.

Martini1991 · 10 Aug 2010 at 15:39

IBT maximum is mainly to stress the IMC imo, due to higher RAM usage, I.E, higher strain on IMC.

At 4129MHZ is nothing on a benchmark, I've done a 3DMark06 at 4172MHZ, and that 4172MHZ was stable for gaming XD.

Hell, even lasted Prime 95 till it throttled with temperatures, which is easy to resolve.

gareth170 · 10 Aug 2010 at 15:42

HeX said:
Not IMHO, i've had prime run for 8hr's solid and LinX/IBT fail in 5mins.

LinX/IBT also heats your CPU up more than Prime.

I usually do 20 loops of LinX/IBT @ 2048MB (or problem size 20,000), if it passes that (takes about an hour) I give it the 8hr Prime Blend treatment.

well for me it's the other way round.

i've had LinX/IBT pass 35 loops @ ~3100MB but Prime on Blend fails after ~2 hours.

HeX · 10 Aug 2010 at 15:45

So the conclusion is, best to run both

But as LinX is the faster test, I say do that first to give an idea of if you are in the ball park for stability

Also LinX gives a good indication if your CPU is throttling (as the GFlops will drop), whereas Prime will give no indication and just plod on regardless.

gareth170 · 10 Aug 2010 at 15:50

HeX said:
Also LinX gives a good indication if your CPU is throttling (as the GFlops will drop), whereas Prime will give no indication and just plod on regardless.

not totally because if background tasks are doing something when linx/ibt are running, the GFlops will drop also.

No_Fear · 10 Aug 2010 at 15:50

good result i am thinking of buying one of those ,good price.

Hagar · 10 Aug 2010 at 17:00

Today I got my thermalright ultra 120 and will start the comparison soon with the 125w version.

I was getting 3400 at 1.28v with poor cooling and airflow, hitting 61C so aborted testing, at 1.18v one core dropped at that OC.

I am using 1.65v on my OCZ gold to get 1600MHz any opinions.

OP, nice results

HeX · 10 Aug 2010 at 18:56

gareth170 said:
not totally because if background tasks are doing something when linx/ibt are running, the GFlops will drop also.

Yeah but you don't run other crap while stress testing, if you are also testing for throttling.

Also if its just normal basic background stuff it won't fluctuate much, throttling you lose a good 5+ GFlops.