Any actual tech-fags here? Do you have any idea if hynix rams

They released the AGESA update, now it's up to motherboard makers to validate memory.

>>59841567
What does this mean?

Is it simply up to mobo manufacturers if they're gonna invest the time to make hynix ram work?

Can anyone elaborate on this?

>>59841515

Can we just have a Ryzen general or something? This is the 5th or 6th thread I see about this.

The /gcg/ exists, so why not a cpu general?

>> 59795428
>>59841515
>>59842941
>>59842706
>>59841945
>>59842362
>>59842459

File: CPUZ bench.png (32KB, 811x407px) Image search: [Google]

32KB, 811x407px

does this actually mean anything?

>>59841567
I installed AGESA update on my ASUS Prime B350 plus yesterday, still couldn't run my Corsair LPX at the rated 3200 16 16 18 36 or whatever. It did better, and I didn't have time to bump the voltage any, but it wasn't stable. It is stable at 2933 15 15 15 36 though.

>>59843605
Have you tried loosening the timings?

Timings really don't make any difference in performance, all that matters is it's mhz.

>>59843654
no, I didn't have time, just set it to what the ram was rated at, and got a blue screen a few minutes into windows having loaded. moved it down to 2933 with tighter timings and ran memtest overnight. I'm going to try again tonight

The key to getting higher RAM speeds for most boards (3200) on Ryzen is either loosening the timings, bumping up the SoC voltage, or both.
People have been getting 3600 with some vodoo magic. Playing with settings and restarting the system multiple times etc. So its not the CPU or memory controler, just the BIOSes need to be polished.
Cant really fault the mobo makers, they've been dealing with Intel stuff mostly for a solid 5 years at least . Ryzen is a new animal with its own intricacies, give them some time and the engineers will get whats going on.

>>59844639
I don't think anyone has gotten above 3000mhz with hynix ram.
Everyone who reported 3200-3600mhz used samsung b-die memory.

Samsung b-die memory is about 30-50% more expensive than hynix ram, so it's a pretty big deal if ryzen gets updates that lets it run on hynix ram.

>>59843654
>Timings really don't make any difference in performance, all that matters is it's mhz

not really, it's just that ryzen processors are more skewed on higher frequency RAM due to the infinity fabric synergy. It's not like timings don't matter at all.

>>59845893
Afaik, it's like 0.5% slower when you double the ram timings.

>>59843363
Half of those are useless shill shitpost threads that should've been deleted on sight.

>>59843605
the update wasn't meant to improve ram compatibility. all it did was lower ram latency by 6ns. granted those modifications amd did to do that broke a lot of patches motherboard manufacturers made for their motherboards to increase ram compatibility. so manufacturers have to once again recreate their patches.

amd agesa update designed SOLELY for increasing ram compatibility is coming in May.

>>59846070
good to know thanks

>>59845893
>>59845940
What timings would you recommend I try to get 3200mhz?

Corsair Vengeance LPX 16GB (2x8GB) DDR4 DRAM 3200MHz C16 (CMK16GX4M2B3200C16) 16-18-18-36

>>59847238
Try very loose, and then work your way down.

Start with like, 22 22 22 50, if it still doesn't work, try a bit looser and so on.
Once you get it working by the big jumps, try to tighten them as low as you can.

>>59847508
is there any kind of rule of thumb for the last number compared to the first 3? like double the lowest value?

>>59847584
No idea really, but something like that.

Watch some youtube videos about it, or read forums. It can be hard to find relevant things, as both youtube and google searching has gone to shit the recent years, where you can hardly even find something related to what you're seraching for.

Honestly, at this point, there' probably no difference between bing and google when it comes to search engine.