Thought it was well spoken and on-target!–I also like the mboard a lot…
I never owned a GB product before now, but I certainly seem to have bought in at the right time! Not having any trouble–3600X–from Zen 1 1600 (bought the X this time around for the higher TDP rating of 95W versus 65W–not sure about that value yet.) Probably will update to an 8-12 core next year, Lord willing …
Information that I’m looking for has been hard to nail down. Do you know what AMD has said–or not said–about the safe max voltage to feed these wonderful CPUs when overclocking? I was able to do an all-core OC to 4.4GHz @ 1.356v on AGESA 1003 (F3 & F4)–I booted Windows, everything looked fine–I ran something to push the cpu (3d Mark, I think) a bit–and wham, down the system went. Garbled my main bios at the time–was able to boot with the backup bios and then restore the main bios. (Great feature! On my x370/x470 MSI mboards it would have been RMA time–dual bios ought to be standard for anyone who overclocks–and, maybe just anyone in general!)
Anyway, I want to treat this 7nm cpu with kid gloves because I know they’ve got to be more fragile–in terms of electromigration vulnerability–than my Zen1 @ 14nm is (it OC’ed 600MHz ROOB just by setting the multiplier and increasing the voltage just a tiny bit–rock solid.) Of course, manually OC’ing automatically turns off boost–which I like. So, I’d like to know realistically what the danger point is for the cpu in terms of voltage–and of course raising the voltage creates heat!–and so on. Any of you guys know or can point me to an AMD resource where max safe voltages for Zen 2 might be listed?
Bear in mind–I’m air-cooled–standard Spire cooler. Also, for me with the Master, AGESA 1003 allows for a boost ceiling of 4.4 GHz with the 3600X, which it reaches exactly–whereas bioses (UEFI, of course) built around 1003AB and 1003ABB lower that to ~4.275 GHz. This is repeatable and easy to demonstrate–looks to me like the AGESA is setting the boost limits. Why? Well, I think you mentioned that lowering the single-core boost seemed to result in better all-core performance with Zen 2–so I don’t know what it’s doing or why. If that’s accurate then I don’t much care about the single-core boost, then.
Here’s what’s interesting–I could manually OC all six cores to 4.4GHz with the AGESA 1003 bioses but with the OC on the latter AGESAs as mentioned above–I couldn’t get much past ~4.3Ghz–sooo…is the boost limit set by the AGESA programming corresponding to the max all-core OC I have been able to get? I feel as though AMD could be much clearer on these topics!
Also, I think I am suffering from “Overclocker’s Mania” these days… I am somewhat amused by my own posts…! Not to mention I have shriveled–become a husk of a man–some people might say I have become obsessed, but I can’t agree, and furthermore–OOOooops!–my foot just fell off! Think I’d better let it rest awhile.
EDIT: 11/21/19 Update
Wow…posted this Aug 5th–seems like last century! Anyway, thought I’d update with a few tidbits on what has changed…
Current Hardware listing:
Windows10x64 v1909, build 18363.476
Latest Adrenalins 19.11.3
AMD 50th Anniversary 5700XT
Aorus Master x570 UEFI bios F10D (AGESA 1004)
Corsair HX-850 PSU (72a x1 12v rail)
AMD Ryzen 5 3600X @ 3.8Ghz & boost
RAM 16GB 2x8GB, Patriot Viper Elite PV416G320C6K @ stock XMP 3200mhz 16 16 16 36 1T @ OC’ed XMP 3733Mhz 18-19-19-19-38-68 1T
LG MultiDrive DVD writer SATA
Boot: Samsung 960 EVO NVMe 250GB (UEFI boot partition)
2x 1TB WD Blue 7200 rpm S3
ST2000DM S3 2TB
ST4000DM004 S3 4TB
BenQ EW-3270U 4k, HDR10 monitor
Now puttering along with AGESA 1004, bios F10D–which I find the best yet, btw. Still no problems with the mboard or peripherals as listed–still very happy with it!
I guess the most exciting thing since August 5th is the ram overclock–don’t know if this is all because of the progress with the AGESA code in these later bios versions, or whether it’s simply the motherboard–but it’s likely to do mainly with the AGESA progress AMD has made, imo, as much as the mboard’s “SuperDuperBrainiacMemoryBoost” feature actually being “super duper”… Anyway, I did not expect to take this 2x8GB ram arrangement and bump the clock to 3733MHz and then test it in P95 memory abuse testing for several hours without a single error! So that is very welcome! I’m still running stock clocks with boost for my CPU, but have been experimenting with PBO lately–still can’t decide whether I think it actually works–but I am seeing three out of my six cores either right on or besting the 4.4GHz max boost of the CPU–and I’m actually very happy with that, atm. Oh, yes–I am using the 1usmus “Ryzen Universal” power plan which seems marginally better than Ryzen balanced–so I’m keeping it for the time being. That’s about it on the update–which may or may not be as interesting as watching paint dry…;)…but I felt remiss if I didn’t come back and add some info! Actually didn’t mean to stay away this long!
@WENDELL: In the event that you happen to peruse this page at some point, I wanted to extend to you a well-deserved compliment. I just watched the clip you and Steve of GN did on the Gigabyte Epyc server platform–the one with the exciting fan start-up noise approximating a 747 taking off… I’m sure you remember it! Seriously, I wanted to make this comment: without disparaging anyone by name, I enjoy your videos because your diction is excellent most of the time and you do not slur your words or do run-on sentences or at times speak so rapidly that what comes out of your mouth sounds like gibberish–a dialect of Elven gibberish, in fact, that is so archaic that I have to spend several seconds running it through my internal translator for Universal Gibberish before I can derive the meaning of what was just said, and so on. Indeed, your speech is clear, precise, and does you credit, especially in making instructive videoclips in which the ability to communicate clearly is of paramount importance. Anyway, the contrast impressed me enough to want to comment on it! Keep up the good work!