I.nfraR.ed

Absolutely insane frequency. Great job!

PBO settings don't matter when in manual overclock mode. Turn SMT off, but leave all cores enabled. Disable all unneeded devices in bios Boot and load windows at whatever all-core multiplier and voltage you can or feel comfortable with., e.g. 4.5GHz, 1.25V. For voltage, there are 2 "manual" modes, one is the full manual where the boards overrides it and you can only change from BIOS or special tools like TurboV. The other manual mode is AMD Overclocking, which can be changed from ryzen master as well. When in windows, you need to find which of your 6 cores can do 5GHz and at what voltage one by one. You may have bad sample and not all cores could do 5GHz. Increase voltage to 1.45V for example and set core #1 (second core) to 5GHz, if it doesn't crash, run superpi with affinity set to that core. If it passes, then you know core #1 is capable of 5GHz. Clock core #1 back to 4.5 and increase core #2 to 5GHz, perform the same test. Do the same for all 6 cores. If the system crashes/freezes at some point, then reboot and continue where you left. You can then test your best cores for minimum voltage as well or if none of the cores can do 5GHz at 1.45V, maybe try 1.5V. That's the general idea, to find a strong core which can do 5GHz. You don't need all the cores at that frequency, just one of them and set the superpi affinity (from Benchmate context menu) to that core, so it will only run at that thread. My CPU can run 5GHz on all cores, but cores #1 and #3 are better than the others.

Have you used the "conditioner" from Stilt?

Haha, ok PS: Just realized all of your scores are back. Nice! I'm happy with the 2nd place for now.

Thanks! This CPU has potential, but either my lack of skills or the board is limiting me. Could not run any other strap and it took me a lot of time to get it 32M stable at that frequency.

Yes, one of Sam's binned CPUs. Didn't put much effort in this, I think ~4.5GHz 32M stable is possible on chiller. FSB-wise I'm not sure how good it is, could run around 605 with x7 multi, but I haven't really tested the board separately or played with GTLs much. Twister is on "Moderate". 1000 6-7-5 is easy with strap 200 and PL5, Twister at "Stronger" at 500 FSB, but max mem freq I could run so far is around 1030-1040, which seems to be the max of this board, at least with stock cooling and with my lack of s.775 skills. Will test it on LN2 some day, should be good for top-5 I think, maybe better.

All this doesn't solve the question "But can it run Crysis at that frequency?", so why bother? I'm all for leaving it as is and don't make things more complicated. CPU-Z doesn't even show what each core is running at, it has to be modified to suit that need and I agree with others that mentioned big.LITTLE and all other complications. PS: I stand corrected, it apparently does now. It shows when I've last "benched" cpuz Technically that's not possible Rest of the cores in the same CCD will have much higher frequency. I think the maximum difference is 800MHz or something like that. Effective frequency is another thing though... but that's also applicable to all-core OC.

Nice work! Next, try PSC/BBSE if you can do ~1100. Also, some of the Gigabyte boards are tricky with hypers and you can't POST at high mem frequency, but you can clock in windows. On my budget GB 970A-UD3 board I have to boot at 1700 and then clock up.

No need for that. Google Drive zips everything inside if you select "Download" on a specific folder. If I want to download a whole folder now it will be inflated and twice the size, because of the additional zip inside.

To me It's just "all in", including upcoming CPUs. It's not a real competition, more like exploring the limits, thus the 1 year lifespan. It can be divided to sub-categories/stages, if there was some sort of a "official" ranking.

Very good score, could not catch that.

Got a decent score, but there's room for improvement, since I had better runs. Spent a whole evening trying to get 4066 running reliably, but it seems my CPU can't handle it for some reason. The better 2x16 kit just doesn't want to run GDM off, so opted for an old 2x8 A0 kit. PS: Top 3 scores are all in our team :p I guess the internal mini-competition helps.

Doesn't seem to make a difference for me. / RANT mode on VTT still doesn't work, anything set manually gets ignored. Or so it seems. 4133/20166 still works at low DRAM voltage, but restarts with higher (I suspect it's the CPU to blame, don't have a spare one). Still WHEA even at 3933. Do these "MemTry It" presets do anything except setting main timings and voltage? They seem to be pointless to me, if that's the only function (in addition to pre-selecting correct DRAM and IF multipliers). I also miss the "Safe boot/MemOK" buttons from Asus. When settings are really wrong, the mem oc recovery mechanism doesn't work and I have to clear cmos. Overall, it doesn't seem faster than my previous board (Asrock B350 K4 flashed to B450 Pro4), just manages more stable voltages and one step on the mem multi up. PS: On a positive note, it might be faster in Pi 32M, but will have to test more. Edit: I've tried very low VTT (0.3V with VDIMM 2v) and it didn't POST, so it probably works.

This is a nice CPU, low FSB hurting the score, though. Keep pushing!

To switch between different channels you need a newer build of ZenTimings where I've enabled that. Just select the corresponding DIMM from the dropdown and it should reload the timings. You can get the latest "public" beta build from my Google Drive https://drive.google.com/drive/folders/108Ux5tv2fMz9OC9SeHR22DmJvfwiNg1Q?usp=sharing

All old dropbox share links are not working anymore. I've shared the folder again (no individual links): https://www.dropbox.com/sh/b702d3y6guzdzyy/AAAq_b0WcaF1fIiQNrUdTjB5a

He has a better one. I literally have no idea about tweaks yet.

Yes, that's because of the hw prefetchers, but on XP and Zen3 haven't experienced the same. Granted my bios lacks those options. Most of the times 3rd run is very slow already and have to restart the system.

I've tried all the tweaks from the old Bulldozer Conditioner (yes, I've disassembled it some time ago and know exactly what it does), also some of my own test tweaks I was working on for the previous zen generations. If anything, one of the registers makes it extremely slow and the whole system gets laggy. Most of the registers are undocumented, the BKDG for 15h has more information than the document for 17h, but there's probably no parity between all the features and a specific bit in a register on 15h might be something else on 17h. Also Zen doesn't support Trailing Bit Manipulation (TBM) instruction set or at least the matching bit is marked as "Reserved" and that's one of the significant tweaks from BDC. Others don't really improve the performance in SuperPi. They had some positive effect on Gen1 last time I tried, but it seems all performance bias tweaks aren't applicable on Vermeer, perhaps nothing significant to tweak anymore with this gen. My modded XP is still working with Vermeer, but is slower than untuned Windows 10. Wazza doesn't help either, so I guess it is game over for XP. I've noticed it was slow with 3000G. It was fine with first gen though, perhaps with Zen+, too. Back on Windows 10 now. PS: Untweaked Windows 7 is also slower for me. PS2: At least now I know which register kills the performance completely and will try to tune the old perf bias options.

@TerraRaptorI don't think so. I will take another look on Zen3, but I could not improve performance on Zen2 much. It had some effect on Zen1 though. The Conditioner uses some MSRs which are also used by the Performance Bias on Ryzen platform, which has no significant effect on the newest gen. Zen doesn't support Trailing Bit Manipulation anyway. I will try to boost performance in SuperPi, but I doubt there's anything significant to gain.

First tries with 5600X, currently on Win 10. Not bad for this board, I think. I can't boot XP anymore with this gen and bios.

Great one. My stick died at similar voltage IIRC.

True, you can't get the real CPU name, it's not fused on the CPU, but is matched to a table depending on the detected features, core, FSB and multiplier. That's why if cache is damaged (or unlocked) it gets detected as something else. Anything running out of stock invalidates the detected model and "stock" sometimes means completely different on different motherboards. It's really easy to cheat with model numbers, but I hope people benching those CPUs are honest. You can basically submit all XP-M Bartons with just one CPU. Even a photo of the CPU doesn't prove anything, except that you have the model, but doesn't guarantee it's the one used for the submission

B350 motherboard running with B450 bios. Wraith Prism cooler.

Oh, you finally did it! Missed that submission. Congrats I will come back for it some day, promise!