Is this a Decent 5950X?

kamikazi

kamikazi

[H]ard|Gawd
Joined
Jan 19, 2006
Messages
1,605
All I've done so far BIOS wise is set the RAM timings to the same as they were for my 3950x. I'm not running PBO. I'm using 64 GB as noted in my signature using the Safe Profile settings from the 1smus calculator. I've run Boost Tester to see kind of what the potential of each core is at stock and I've run CB20. Anyway, here's a screenshot of Hardware Info after a few CB20 runs.

I've seen conflicting info on line for stock CB20 scores where it seems some are low in reviews and others look like PBO was applied, that's why I'm asking. I will say that this chip seems to run 2 to 4 degrees C higher at idle than the 3950x. I may try one remount just to see if I screwed up. I am using that Der8aur Ryzen OC bracket. The CB20 scores were run with no monitoring app in the background.

5950x 05-27-21.jpg
5950x R20 unmonitored 05-27-21.jpg
 

Attachments

  • 5950x 05-27-21.jpg
    5950x 05-27-21.jpg
    276.9 KB · Views: 0
Anandtech got 10,096 multi and 644 single. Seems like some decent deltas between testers.
 
Gonna have to remount. I hit 78 C playing Dragon Age Origins using the same settings and fan profile from my 3950x. Fans were constantly ramping up and running loud. With the 3950x, they pretty much stayed constant at low noise. I have the curve set for 40 percent at 55 degrees and 100 percent at 75 degrees.

I hope I didn't get a bad chip like one of the 3900x chips I binned. One of those ran almost 10 C hotter than the other. I remounted that cpu probably 5 or 6 times before I gave up.
 
Here's a shot from my initial run with my 5950x on an Asus x570-i, Gskill 16gb 3200mhz cl14, Arctic Cooling 280mm AIO. I believe this was with PBO enabled.


CB20.jpg
 
I have a 5900x and it will barely boost to 4.9ghz on 2 cores. I believe they are being really loose with some of the specs of these chips due to supply constraints.
 
I wish either of my 5950Xs were as nice as that one. One chip has 2 cores that will boost to 5 ghz on the rare occasion at start up, the other has none. Neither has a second chiplet with a core that will hit 4.9.

CB20 single core on one can't break 627 on the other it's 612. Ryzen CTR tells me one of my chips is a "Golden Sample" because it clocks decent at low voltage, but neither can do over 10000 Multi without PBO. Really got rooked this round for sure.

I'd say yours is a keeper.
 
Xero717 said:
I wish either of my 5950Xs were as nice as that one. One chip has 2 cores that will boost to 5 ghz on the rare occasion at start up, the other has none. Neither has a second chiplet with a core that will hit 4.9.

CB20 single core on one can't break 627 on the other it's 612. Ryzen CTR tells me one of my chips is a "Golden Sample" because it clocks decent at low voltage, but neither can do over 10000 Multi without PBO. Really got rooked this round for sure.

I'd say yours is a keeper.
Click to expand...
Seems these chips are all over the place. I need to figure out how to tune it down on old games where I'm not gpu limited as well as single core stuff. I fired up Destiny 2 and didn't get anywhere near the temperature spikes and fan noise as I did with Dragon Age. Maybe super sampling on the gpu could help.

I've read up and lots of people are seeing the same issues with the 5950x concerning high temp spike while doing mundane stuff. I followed on rabbit hole where they recommended turning off Clock Boost and PBO in BIOS, then install Asus AI Suite 3. Let it auto-clock your cpu, then go back to BIOS and turn on Clock Boost and PBO. I wish I didn't waste my time. Turns out, all that does is set up a crummy all core overclock which avoids super high single core speeds and heat. The writer of the comments didn't mention that. Hopefully, that little Asus experiment didn't degrade my chip. I'm typically only seeing 3 cores at 5 GHz or better now, not 6 like above. However, I did also slightly changed my cooling from using the Der8auer Ryzen OC bracket to just mounting the waterblock as normal. Looks like my idle temps are a degree or 2 higher. Who knows.

I did run the 1usmus CTR utility and just did the diagnosis. It said I had a silver sample. I need to look into all that curve optimization stuff in the BIOS now.
 
i need to dig into curve optimizer still seeing spikes in voltages up to 1.525 randomly would be nice to max the voltage to 1.35 1.40 if posssible.
 
Well, one surefire way to avoid spikes in temps is to set the cpu to thermal throttle at 70 or 75 degrees. By default, it's 95 degrees, right? When I ran that Asus tool, it also ran fan expert to set up fan speeds. It was checking for silent mode or something and stopped the fans. I hit the stop button at 93.5 degrees. I wasn't going to trust it to kick in at 95.
 
kamikazi said:
Seems these chips are all over the place. I need to figure out how to tune it down on old games where I'm not gpu limited as well as single core stuff. I fired up Destiny 2 and didn't get anywhere near the temperature spikes and fan noise as I did with Dragon Age. Maybe super sampling on the gpu could help.

I've read up and lots of people are seeing the same issues with the 5950x concerning high temp spike while doing mundane stuff. I followed on rabbit hole where they recommended turning off Clock Boost and PBO in BIOS, then install Asus AI Suite 3. Let it auto-clock your cpu, then go back to BIOS and turn on Clock Boost and PBO. I wish I didn't waste my time. Turns out, all that does is set up a crummy all core overclock which avoids super high single core speeds and heat. The writer of the comments didn't mention that. Hopefully, that little Asus experiment didn't degrade my chip. I'm typically only seeing 3 cores at 5 GHz or better now, not 6 like above. However, I did also slightly changed my cooling from using the Der8auer Ryzen OC bracket to just mounting the waterblock as normal. Looks like my idle temps are a degree or 2 higher. Who knows.

I did run the 1usmus CTR utility and just did the diagnosis. It said I had a silver sample. I need to look into all that curve optimization stuff in the BIOS now.
Click to expand...
Its likely that Asus software made changes in your bios settings. I would reset your bios to defaults and start over on your bios settings. And then as you reach points where you have things tweaked a certain way which you don't want to lose --- save a configuration profile (often referred to as an overclocking profile). That way, if something gets messed up as you continue to tweak things, you can easily go back to that milestone. instead of starting from scratch. Nowadays, you can even backup those profiles to a thumb drive.
 
chameleoneel said:
Its likely that Asus software made changes in your bios settings. I would reset your bios to defaults and start over on your bios settings. And then as you reach points where you have things tweaked a certain way which you don't want to lose --- save a configuration profile (often referred to as an overclocking profile). That way, if something gets messed up as you continue to tweak things, you can easily go back to that milestone. instead of starting from scratch. Nowadays, you can even backup those profiles to a thumb drive.
Click to expand...
I did load my profile back on there, but I didn't reset it to defaults first. That's a good idea. I did make sure I saved a profile when I had it the way I wanted it.

I also changed my windows power option from the 1usmus power setting to just Windows Balanced since that is supposed to work with Ryzen 3.
 
kamikazi said:
I did load my profile back on there, but I didn't reset it to defaults first. That's a good idea. I did make sure I saved a profile when I had it the way I wanted it.

I also changed my windows power option from the 1usmus power setting to just Windows Balanced since that is supposed to work with Ryzen 3.
Click to expand...
If you had a profile saved before you installed the Asus software, you don't need to reset to defaults first, before you re-load the profile ;)
 
Just upgraded from a 5900x to a 5950x this past week. Running memory at stock settings (32Gbs of DDR3600 @ CAS 14) with Infinity Fabric set to 1:1, PBO enabled and Ryzen Master configured via Simple View with "Auto OC".

No real big changes when it comes to single core performance (sort of expected that), but happy with the nice uptick in multi performance with the 5950x pushing the extra cores:

1624057842409.png
 
Supercharged_Z06 said:
Just upgraded from a 5900x to a 5950x this past week. Running memory at stock settings (32Gbs of DDR3600 @ CAS 14) with Infinity Fabric set to 1:1, PBO enabled and Ryzen Master configured via Simple View with "Auto OC".

No real big changes when it comes to single core performance (sort of expected that), but happy with the nice uptick in multi performance with the 5950x pushing the extra cores:

View attachment 367397
Click to expand...
Nice, I got about that score today using PBO, motherboard power limits, and custom curve with each core set to -10 mv. However, Windows will blue screen on boot every other time. No problem running CB20 or Shadow of the Tomb Raider though. CB20 will push temps up to 90 though. I have a lot of tuning to do.
 
It probably helps that I have way overkill when it comes to my cooling. :)

My CPU idles at 29C and I never see it get above 60C under extreme loads.

For stability testing, I'm use OCCT (ver 8.2.1). If the system can handle AVX2 on a small data set in extreme mode with a steady load for an hour without throwing a single WHEA error, then I consider it fully stable.

I did try some G.Skill 4000 CAS 14 rated RAM just for fun (was just released a few weeks ago) to see if it I could eek out a bit more performance... but that speed/timing with a 1:1 Infinity Fabric was just too much for the 5950x to handle. It would boot, but I was getting way too many WHEA errors and dialing it back from its stock memory speed/timings seemed silly. I returned it and went back to my trusty G.Skill DDR4 3600 CAS 14 which is solid as a rock with Infinity Frabric set to 1800 (1:1).

I may play with a more advanced OC at some point down the road (Cores/Mem), but then again, it's stable now and performs well, so it doesn't really seem worth all the effort and monkeying about. :D
 
Supercharged_Z06 said:
It probably helps that I have way overkill when it comes to my cooling. :)

My CPU idles at 29C and I never see it get above 60C under extreme loads.

For stability testing, I'm use OCCT (ver 8.2.1). If the system can handle AVX2 on a small data set in extreme mode with a steady load for an hour without throwing a single WHEA error, then I consider it fully stable.

I did try some G.Skill 4000 CAS 14 rated RAM just for fun (was just released a few weeks ago) to see if it I could eek out a bit more performance... but that speed/timing with a 1:1 Infinity Fabric was just too much for the 5950x to handle. It would boot, but I was getting way too many WHEA errors and dialing it back from its stock memory speed/timings seemed silly. I returned it and went back to my trusty G.Skill DDR4 3600 CAS 14 which is solid as a rock with Infinity Frabric set to 1800 (1:1).

I may play with a more advanced OC at some point down the road (Cores/Mem), but then again, it's stable now and performs well, so it doesn't really seem worth all the effort and monkeying about. :D
Click to expand...
I thought my cooling was overkill and I idle much higher than you do. How in the world do you get that CB20 score and keep cpu at 60 and under? Which sensor are you reading?
 
OK, I guess it is closer to a 67-68C max. I fired up HWMonitor to have an alternate look during a bench/stress run - seems it peaks closer to 68C and idles at 32C. Still not too shabby. I think I was looking at the Dark Hero CPU temp sensor previously. Probably should be using the CPU Die (average) from the 5950x that's in OCCT instead for a more precise read of CPU temp.

1624066616566.png


Edit: Just did a few multi CB20 runs and it peaked at 75C in HWMonitor. (The single core CB20 run was where I was probably noticing the ~60C max I bet)
 
Last edited:
Supercharged_Z06 said:
OK, I guess it is closer to a 67-68C max. I fired up HWMonitor to have an alternate look during a bench/stress run - seems it peaks closer to 68C and idles at 32C. Still not too shabby. I think I was looking at the Dark Hero CPU temp sensor previously. Probably should be using the CPU Die (average) from the 5950x that's in OCCT instead for a more precise read of CPU temp.

View attachment 367425

Edit: Just did a few multi CB20 runs and it peaked at 75C in HWMonitor. (The single core CB20 run was where I was probably noticing the ~60C max I bet)
Click to expand...
thanks for actually validating your temps. I see all the time, people claiming pretty excellent temps relative to their cooling solution and I'm just like....mmhmmm....... >_>
 
You must log in or register to reply here.
Back
Top