subaruwrc

I bought 2 other G3258 chips today. Both banana.. L421B829 #0644 VID 1.116v 4.1ghz 1.2v cinebench 15 ok - 4.2ghz fail L421B829 #1630 VID 1.092v 4.2ghz 1.2v cinebench 15 ok - 4.3ghz fail 4.5ghz cb15 needs 1.4v

I removed the ihs from 2 cpu-s with different benefits. 1st was a L417B973 malay chip which did 6.1 max valid on stock configuration under ln2. Didn't bother to check air results with that one. Then I removed the ihs and changed the paste underneath with gelid gc extreme. Benefit is 200mhz, 6.3 valid and 6.1 superpi 32 but thats the max. Another chip is a 3419 costa rica which got delided and the ihs polished in addition because it was concave but the benefit is a short 100mhz in max validation. previous 6.1 is now 6.2 with 1.825v 2.4v vcin,no voltage scaling above this. I got a little mad and tried the chip with on die ln2 cooling but the results were far worse as the chip did only 5.9ghz max validation with the ihs removed. Chip is alive but useless. I hoped to gain at least 300mhz cause it was a hot chip and got bad ihs but no luck. Anyway it did 1.54v superpi 32m at 5ghz on air. 5170mhz 1 core valid with 1.55v and 5.2ghz with 2 cores with 1.54v after changing TIM.

In case of 1 core loading. I think if the benchmark ran at 4.9ghz it wont be only a couple of points better than the previous 1184marks result with lower uncore. I know this guy very well,he is not a cheating kind of guy.. In addition if you search for real 4900mhz results those are in the 1260+ marks area up to 1280.

even with more vcore like 1.7-1.8v ? why so low voltage ?

You are on fire,keep pushing! I love your work.

sick setup !

G3258 tested again today. 3419B307 1,191v 4500 wprime 32m,cinebench 15 1.240v 4600 wprimw 32m cinebench 15 1.417 1.9v vccin 4900 spi 32m 1.55v 2.2v vccin 5000 spi 32m ... 2nd chip L417B973 strange. vid 1v ... 1.2v 4400 boot 1.35v barely 4800 valid

standard question: what batch? after several banana g3258 maybe its time for me to buy a decent 47x0k ..

newer batches seems better on average.

thanks man. I try to imrpove it soon!

real men use CL5 !

6.3 32m with 1.7v in a world where i cant reach 6.3 cpuz at 2v... nice work man

sick

memory settings are important in this test. you are running 2400 c10 while nacho arroyo runs 2800 c9. do the math.

grat, keep pushing!

Simply awesome effort there !

this is my favourite, killer setup, awesome effort, keep pushing!

awesome effort man!

kill it with fire.

look for lowest possible voltage needed for 5ghz superpi 32m at ambient cooling. these numbers are not comparable.

yes i totally understood that and i said what i said because the 5-6 chips i tested was a lot worse than this and im a bit jelaous 6.3 max valid, not even 6.1 max 32m from any of my chips. now keep pushing and go for a killer 32m time!

i have a kidney to sell if you interested ...

L251A899 4670K ES - 1.35v 4800 2c/2t 32m Ok, 1.575v 4900mhz FAIL. oh my ..

looks like a good bin.