![]() Run AIDA64 again and see what you have gained. When you manage to boot and get in windows first thing is to open RM and see that you have the correct settings. Once all set (double checked) and before save and exit I suggest to save this profile preferably to a usb stick. This disables the power savings of the SOC/IO die and can help with IF/UMC OC (can raise the avg power consumption of SOC/IO by a few watts, but nothing really important)įeel free to ask further questions and/or clarifications. You also need to be sure about the matching speeds of DRAM and UMC so set the UCLK DIV1 MODE to: UCLK=MEMCLKĪlso you can set the SoC/Uncore OC Mode to Enable. ![]() Under AMD CBS and XFR Enhancement (settings page) in BIOS you will find the IF speed ( FCLK set it to 1800MHz) Max recommended safe 24/7 mem voltage especially for b-dies is 1.45V and some users do even 1.5V with active cooling upon modules. Mine works with 1.4V (DRAM voltage CH A/B) for these kind of settings. The 1.35 recommendation for mem voltage is really for high end/quality B-dies and ours are not. Be aware that the timings order is different in BIOS. Set the mem multi to 36 and timings in the Advanced Memory Settings. All these settings are in the Tweaker page of your BIOS. You dont have to set all the settings provided here but just those in blue squares for start. When you get the current performance you can proceed with DRAMcalculator like this Run them 1 or 2 times each and save the numbers. You can find the first pic memory/cache benchmark right above the Start button in Tools drop down menu. If you dont want to pay for it then you can run the individual 4 tests for memory read/write/copy/latency like this. Your bandwidth could end up higher than mine and latency lower due to higher threads and speed of your CPU.īe aware that AIDA64 is a paid benchmark. You can run "AIDA64 cache and memory benchmark".įirst raw is the important about mem read/write/copy bandwidth (higher) and latency the most out of all (lower) Major factor about RAM OC for Ryzen is the UMC (Unified Memory Controller) of the CPU and I bet yours is better than mine cause R9 3900X is a higher quality chip overall than R5 3600.Īs i thought the modules are 1 ranked (says so the "Memory Module" -> "Organization" on Thaiphoon) meaning that all 8GB is on 1 side of the modules like most 8GB sticks.įirst you need to get a starting point for your DRAM performance and how is it now. Ok the ICs (actual memory chips) are a bit worst than mine but it doesn't matter all that much. Ryzen DRAM calculator has everything you need, maybe check the V2 profile instead for a more relaxed set of timings. Throw a little bit extra DRAM voltage at it and you'll likely be able to comfortably get a little frequency and tighter timings out of it. I described Corsair as "poorly-binned" B-die, but it's still B-die. On the other hand, if your chip is like mine with Core 0 and the best core at opposite ends of the silicon, the scheduler will feel a bit torn at times. ![]() Your Core 0 quality makes it easy, because as far as the scheduler is concerned, there is literally no reason ever to migrate load out of the first CCX unless more than 3 cores are needed, which is probably why you see lower freqs and usage on the other 9 cores. Seeing as your two best cores are in the first complex, it's not hard to see why the scheduler would want to keep most of the load in there. For you, a CCX is populated by 3 active cores. If you're using 1usmus or 1909, the scheduler tries its best to keep lightly threaded loads within the same CCX for latency purposes. Better for AMD to set an unattainable ceiling than one that will actually limit performance. ![]() It's like the theoretical power draw limits under manual OC - when will you ever reach the limit of 1000W of power draw on anything? It's pointless to fret over, and has no bearing on what you actually get.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |