He lowered the voltage because he said the CPU was running hot - it was idling at 40c under stock settings - which is actually hot for any cpu...but amd says it’s normal so leave it be - your cpu has a warranty so who cares if it’s idling at 1.5v as long as it doesn’t die that’s fine
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
AMD Zen 2 (Ryzen 3000) - *** NO COMPETITOR HINTING ***
- Thread starter Jamin280672
- Start date
Associate
Not so easy in Australia. To go through RMA most of the stores will test first before sending back to manufacturer, if they conclude there's no issue (without testing beyond sticking in a mobo and testing idle temps normally) then it never gets to back to the manufacturer to do a proper test.
I was quite lucky I had two 3700X's at the same time, one was hitting 90C under any load even when locked to 3.6GHz while the other at 4.4GHz would hit 82C under load. Immediately could see the CPU was duff and returned. When people buy solo though, that could cause many days/weeks of troubleshooting trying to find out the CPU was knackered.
AMD definitely has the more "beta" feel about it.
The problem I have with jayz video tho is he hasnt swapped out board or cpu to confirm if either is faulty, he has just assumed its a bios issue. I expect he wont swap out unless he gets replacement free kit as well, as well thats what reviewers deserve right.
I am glad he made the video tho, this stuff shouldnt be hidden.
I agree with him on the stock voltages.
Associate
- Joined
- 21 Sep 2018
- Posts
- 895
It took another tech reviewer to point out his flaw.
Associate
- Joined
- 2 Jan 2019
- Posts
- 617
Anyone else with an ASUS X570 mobo on ABBA AGESA tried using the "OC Tuner" button in their BIOS?
I flashed the 1201 BIOS on my Asus Tuf Gaming X570-Plus (WiFi) yesterday, then accidentally clicked it having changed nothing else (including not even setting DOCP on the RAM). The system quickly reboot with a 4.1GHz base clock at 1.1v.
I ran Cinebench R20 MT and got higher scores than previously where it would also run Cinebench at 4.1GHz but at 1.33v. Of course, it wouldn't then boost beyond 4.1GHz on a single core.
Naturally, I figured that I could do better myself so I set the clock targets to 4.275GHz all core at 1.1v. Cinebench ran perfectly fine, and gave a MT result that was almost identical to when I previously ran 4.35GHz all-core at 1.425v in the past.
Of course, the ABBA AGESA was primarily supposed to be about achieving advertised boost clocks. Having not actually tested this on my accidental "OC Tuner" mess-around, I decided to reset the BIOS to defaults.
Cinebench MT runs remained the same at around 4.1GHz at unnecessarily high voltages. ST did boost higher than I'd seen on all previous BIOSes (other than release BIOS), and it ran Cinebench at around 4.325GHz, sometimes higher and other times lower. For the first time I got a score in excess of 500 at 502.
In HWINFO64 I also saw that at least 2 of the cores exceeded 4.4GHz in the, admittedly very brief, time that I had it running. Previously it would only get to 4.325GHz at best.
The question I have for myself is whether I feel it worthwhile to let it operate normally, or to utilise the OC Tuner function but use Ryzen Master to set it to target 4.275GHz. Do I feel it is worth the extra performance at default over reduced temps/power using OC Tuner?
As an added extra I also tried increasing the voltage in OC Tuner mode with increased clocks. I was able to run Cinebench MT at 4.375GHz for 1.325v, having previously only been able to run 4.35GHz at 1.425v. I set my new highest score of 3948 against 3874 in the past (3875 at 4.275GHz at 1.1v).
So two observations from my less than perfect testing merhodology with ABBA AGESA; higher clocks (at default and OC Tuner mode), higher scores (at default and OCTuner mode). A third observation was that my temperatures were (for the first time) very close to the 95C thermal limit under some of the test conditions.
I flashed the 1201 BIOS on my Asus Tuf Gaming X570-Plus (WiFi) yesterday, then accidentally clicked it having changed nothing else (including not even setting DOCP on the RAM). The system quickly reboot with a 4.1GHz base clock at 1.1v.
I ran Cinebench R20 MT and got higher scores than previously where it would also run Cinebench at 4.1GHz but at 1.33v. Of course, it wouldn't then boost beyond 4.1GHz on a single core.
Naturally, I figured that I could do better myself so I set the clock targets to 4.275GHz all core at 1.1v. Cinebench ran perfectly fine, and gave a MT result that was almost identical to when I previously ran 4.35GHz all-core at 1.425v in the past.
Of course, the ABBA AGESA was primarily supposed to be about achieving advertised boost clocks. Having not actually tested this on my accidental "OC Tuner" mess-around, I decided to reset the BIOS to defaults.
Cinebench MT runs remained the same at around 4.1GHz at unnecessarily high voltages. ST did boost higher than I'd seen on all previous BIOSes (other than release BIOS), and it ran Cinebench at around 4.325GHz, sometimes higher and other times lower. For the first time I got a score in excess of 500 at 502.
In HWINFO64 I also saw that at least 2 of the cores exceeded 4.4GHz in the, admittedly very brief, time that I had it running. Previously it would only get to 4.325GHz at best.
The question I have for myself is whether I feel it worthwhile to let it operate normally, or to utilise the OC Tuner function but use Ryzen Master to set it to target 4.275GHz. Do I feel it is worth the extra performance at default over reduced temps/power using OC Tuner?
As an added extra I also tried increasing the voltage in OC Tuner mode with increased clocks. I was able to run Cinebench MT at 4.375GHz for 1.325v, having previously only been able to run 4.35GHz at 1.425v. I set my new highest score of 3948 against 3874 in the past (3875 at 4.275GHz at 1.1v).
So two observations from my less than perfect testing merhodology with ABBA AGESA; higher clocks (at default and OC Tuner mode), higher scores (at default and OCTuner mode). A third observation was that my temperatures were (for the first time) very close to the 95C thermal limit under some of the test conditions.
Soldato
The Ryzen 9 3950X is delayed until November:
https://www.anandtech.com/show/14895/amd-next-gen-threadripper-and-ryzen-9-3950x-coming-november
https://www.anandtech.com/show/14895/amd-next-gen-threadripper-and-ryzen-9-3950x-coming-november
Caporegime
Well... No! this shows just how much of an idiot he is...
So he gets into Windows, complains about 1.5 volts idle and invites his viewers to watch how high the temps will go when he runs Cinebench, 63c and the volts dropped to 1.26v during the run. its pretty obvious he was not expecting that, like an utter NOOB he looked at the idle volts, thought he knew some thing he didn't even begin to understand, the 1.5v Idle, which is completely normal, and panicked, like someone who knows 1.5v is high for an Intel but knows absolutely nothing beyond that.
Most of the people in here know far more about CPU's than he does.
https://youtu.be/XdxyTWSeRx4?t=359
Like this ^^^^ guy he doesn't understand how the boost algorithm works. And i'm sick of explaining it.
Last edited:
Soldato
msi in comments.
Caporegime
Soldato
I thought that was amusing when Jay2Cells was on about the 1.5V at idle, then loaded up R20, expecting the temps to hit eighty or ninety using megavolts, then he was like "oh, 1.2V, oh, 63C". 
Soldato
Considering they are supposed to be "tech" youtubers, you'd have thought they would actually learn how an entirely new CPU architecture actually worked and behaved before publishing videos on them!
Everyone just assumed they'll work like Intels, and can't possible fathom it might actually work differently and require different voltages to work than the competition.
Everyone just assumed they'll work like Intels, and can't possible fathom it might actually work differently and require different voltages to work than the competition.
Soldato
People on high horses need to own up when they screw up or are a bit out of their depth. One of the issues of always fanboying to one side (as everyone has their preferences) is make sure you spent time and learn about the nuances before flaming the product.
Associate
- Joined
- 2 Jan 2019
- Posts
- 617
Update on my 3600X.
Running at optimize defaults, DOCP, and +200MHz Auto-OC, with ABBA AGESA I now see 3 of my 6 cores exceed the 4.4GHz on the box, a 4th hitting the x44 multiplier, and the other 2 hitting a 43.75x multiplier. My mobo has bus clock of 99.8MHz, so the x44 multiplier is more like 4.391GHz.
That's a clear improvement over an earlier AGESA that I was running (it wasn't the ABB AGESA though).
CPU Package Power rarely exceeds 65w, though I have seen 99w under load. Temperatures seem pretty reasonable, and the fact it can now run CBR20ST at (above) 4.325GHz, which the previous BIOS couldn't achieve under any auto setting, is a good sign. Still only hitting 505 in Cinebench though, so not quite the 510 that I saw from one lucky 3600 owner in here.
Running at optimize defaults, DOCP, and +200MHz Auto-OC, with ABBA AGESA I now see 3 of my 6 cores exceed the 4.4GHz on the box, a 4th hitting the x44 multiplier, and the other 2 hitting a 43.75x multiplier. My mobo has bus clock of 99.8MHz, so the x44 multiplier is more like 4.391GHz.
That's a clear improvement over an earlier AGESA that I was running (it wasn't the ABB AGESA though).
CPU Package Power rarely exceeds 65w, though I have seen 99w under load. Temperatures seem pretty reasonable, and the fact it can now run CBR20ST at (above) 4.325GHz, which the previous BIOS couldn't achieve under any auto setting, is a good sign. Still only hitting 505 in Cinebench though, so not quite the 510 that I saw from one lucky 3600 owner in here.
Caporegime
Looks promising, if its the same 3600 owner that i'm thinking of he did have PBO +200 to get that score, i'm getting 486 at stock, 1003ABB. I'm waiting on ASRock to release my new shiny.
Associate
- Joined
- 2 Jan 2019
- Posts
- 617
The +200 doesn't seem to do much for me. Whilst the target is set for 4.6GHz, it doesn't go anywhere near that level.
Personally, I couldn't get mine to do a 4.4GHz Cinebench ST run, which is bizarre because 4.375GHz all-core worked fine when set manually, and it's at 4.325-4.35GHz by itself.
The cores themselves are good for 4.5GHz on a light stress test like the one in Ryzen Master itself.
Personally, I couldn't get mine to do a 4.4GHz Cinebench ST run, which is bizarre because 4.375GHz all-core worked fine when set manually, and it's at 4.325-4.35GHz by itself.
The cores themselves are good for 4.5GHz on a light stress test like the one in Ryzen Master itself.
Caporegime
Cinebench is a higher stress load, so with an overclock it will only boost higher than it otherwise would for that load, so if in Cinebebch at stock you might boost to 4.15Ghz, +200 might get you 4.35Ghz, or a lightly threaded game maybe 4.3Ghz stock and 4.5Ghz with +200Mhz.
To me it looks a lot like GPU Boost 3 on my Pascal GPU, 1911Mhz out of the box, i set +100Mhz gives 2000Mhz or slightly under, however as it heats up it gradually drops the Mhz, by about 3Mhz increments, if i have been gaming for a while it might go from starting at 1997Mhz to 1963Mhz.
If you open up Ryzen Master and run a high stress load keep an eye on the Mhz, it might start at 4.4Ghz and gradually fall to 4.35Ghz in 1, 2 or 3Mhz increments, i don't know how it does this with 25Mhz multiplier but it does.
To me it looks a lot like GPU Boost 3 on my Pascal GPU, 1911Mhz out of the box, i set +100Mhz gives 2000Mhz or slightly under, however as it heats up it gradually drops the Mhz, by about 3Mhz increments, if i have been gaming for a while it might go from starting at 1997Mhz to 1963Mhz.
If you open up Ryzen Master and run a high stress load keep an eye on the Mhz, it might start at 4.4Ghz and gradually fall to 4.35Ghz in 1, 2 or 3Mhz increments, i don't know how it does this with 25Mhz multiplier but it does.
I'm glad I wasn't the only one who watches Jay's video and wondered what he was talking about with the voltage and temps.. I had thought to myself that AMD already this behaviour, why is he making it out like its an actual issue? Fair enough the board was having a brain fart but immediately blaming this on AMD and not MSI was an odd one (I say this as the problem appeared to be board / bios specific rather than systematic). Maybe he should have geared the video towards fixing the issue and making more of a "tutorial" on what to do if you do encounter such a problem.
That said I do enjoy his build series where he goes more in depth into the modding side of things. I really enjoyed the recent Destiny 2 themed build he did for Phil's friend. No attempt at technical stuff just modding.
Stability wise I have had no issues with my X570 Aorus Pro / 3900X system, barring the F6a ABBA quirk which prevents normal boot - functionally in windows it is fine. Additionally my wife's "old" B450i / 2600 has been rock solid since day one. Didn't even bat an eyelid when I gave it new RAM and it immediately picked up the XMP profile without even setting it in the bios.
I *did* have issues with Ryzen 1st gen but from what I have experienced things have got a lot better.
That said I do enjoy his build series where he goes more in depth into the modding side of things. I really enjoyed the recent Destiny 2 themed build he did for Phil's friend. No attempt at technical stuff just modding.
Stability wise I have had no issues with my X570 Aorus Pro / 3900X system, barring the F6a ABBA quirk which prevents normal boot - functionally in windows it is fine. Additionally my wife's "old" B450i / 2600 has been rock solid since day one. Didn't even bat an eyelid when I gave it new RAM and it immediately picked up the XMP profile without even setting it in the bios.
I *did* have issues with Ryzen 1st gen but from what I have experienced things have got a lot better.
Caporegime
I enjoy Jays channel too. I find it entertaining, but he did seem to be ranting for no reason whatsoever at AMD. A poor effort.
Soldato
This has made a big difference for me, and I see Asus have released it officially now.
- 4.4GHz boost clocks being hit
- Vcore fluctuating as it should
- No more clear CMOS (so far, so good).
Soldato
Yeah I'm really happy with it. Asus stepping up their bios game is good to see.