chew*

10200 gb3 done........

it's not the cpu....it's more the memory......my cpu's are not as good as people think they are ......my memory however is alot better than people think. I can literally put 1 stick in at a time on a profile and c5 a vast majority of them at just 10600.

The Lords work is never done........PM me for input to fix the board This chip is fclk challenged but more so on currnet bios in this board.

yah not bad for stock cooler., Slot closest to cpu., get off the apu it's a headache. the profiles subs are "tight" so like i said will need good M. I can more than likely pull in primaries just wanted a reliable baseline to work from,

https://valid.x86.fr/p7h3cw Profile to do it. decent not great decent cpu needed....good M die necessary. 10600 asrock hdv

10k geekbench profile. works on 3 out of 4 of my cpu's works 9800 on the dud. https://cdn.discordapp.com/attachments/622285083591573506/1210091102518911016/10kgeekbench?ex=65e94c29&is=65d6d729&hm=d5c47cc0b897dd66c3848b27de8ec098bd77a6eb0a2fee7a46eb4a90cac36410&

a die tune for 9200.

subs bad for M on AMD/tachyon. This is the way......and it's with a 8600g which is slower in geekbench Cheers

Its not A wasn't chasing a "score" was chasing a frequency. 10K done

10600 done https://valid.x86.fr/lzv9uy

Where all the tachyons went......

https://valid.x86.fr/ax865g

Just screwing around with some M die.....testing to see where AMD's imc falls off in the stable department

it's not going to get rid of CB sergmann. This is a test bios with unlocked SOC but as you can see......and at cookies request I locked it at 1.3 just to show it's not needed. Honestly for 1/2 mode high vsoc doesn't do much. You only need high soc for 1/1 due to soc feeding uclk which is much higher 1/1. 8800 is easy at 1.3 soc if the imc is capable. You can see my soc voltage is set to 1.3 for this. This is max boot to bios.....I've already lost sata at 109.45........I lose the vga over this point. This is like 2200+/2200+ uclk/fclk Here's another cpu on just air.....( temps do matter ) and this one will probably scale if I pulled it down to 21c like the one above.....

We have my 5500 run for comparison anyway....... He is 3.4 secs behind.

x3d cpu's are 10 secs faster cpc.........the vcache is just faster in 2d also.....even in py prime etc etc......

Pretest for 6gig run non v cache 5.5 so you can calculate where that effeciency is going to land edited because screens dissapeared

I was going to play the whip out 1 second faster at a time game but since doms bored......i'll shove this one up off thumb drive for now. Still on the dirty/broken OS. No need for tachyon or a tuned OS or tweaks just yet Processes running shown for transparency. I'm not that guy that claims bloated OS and has a stripped to hell OS

Np man. You know I don't care about spare hardware lying around need it lmk.

Want my Asus?

Def not at ambient. I think the benchable 6600 imc 7950x might have 1 core that could do 6 with a cold radiator/Ac on. Yah I was not sure if your able to get 1 core to run at like 5.6-5.7 via eclk on your 7800x3d dom. If so there's a slight chance that a barely stable 6 gig run sub zero might be doable. Need like 123.7xx Eclk pbo +30 which is already dangerous vcache volts with a -200 boost so cpu does not do stupid shit and work off the 48.50x multi. Even then it will probably not be a effective clock 6 gig pass due to the weird discrepancies i've seen with effective clocks vs actual. I'll think about it I guess when i grab another spare x3d. Probably be easier on 7950x3d with it's 200 higher fmax. I need to get some issues sorted on giga before i even attempt it. HV mode Mem clocks with Async mode currently do not play well together.

Yah it was really odd. The Vrm would just "quit" no magic smoke. High loads like wprime 1024m board would just power down and cpu would be dead. Happened on mad once happened to us twice and then we excluded wprime 1024M from our thuban testing on the boards from that point on. Gen 2 or refresh boards after 990 launched are probably best suited for load capability of BD/vishera.

14 secs apart vcache to non vcache then and overall improvement for x3d of 4 seconds. I guess not horrible considering ddr4 Imc was at it's peak and ddr5 imc is infantile at best. Looks like 10 sec improvement from non vcache as I think 5:36 is very realistic from my quick tests. Literally was knocking on the door to 5:36 Dom how close can you get to 6 gig on current cooling on your x3d? DI which should be a easier temp to deal with coldbug and the tuning for 32m is probably worth about 200-300 mhz at a "fixed" voltage so curious if a 6gig run would be even realistic/possible.

I would suggest not trying to OC it to hard. Myself and mad22 (GALAX) found in competition that the vrm was less capable than previous gen (790 gigas) boards and we both managed to kill cpu's and boards with just thuban. BD/Vishera are capable of pulling even more than thuban.

Honestly I have a 5800x and 5800x3d and the difference from that ( just in general from 1 cpu to the other ) was so negligible from 3000 series it honestly did not interest me. I only show up when there are significant changes to hardware.