|
How many of you run a modern(ish) Ryzen?
They seem to run a little hot. 65C while stress testing, but I read somewhere these chips are good for up to 90 degrees or so.
I know nothing about AMD. Am I totally off base here thinking my chip is actually running pretty cool for this series of chips?
Real programmers use butterflies
|
|
|
|
|
65 C while under stress is very good... The 9-line is graded up the close to 100 C, the 7-line is around 85 C...
Probably with age you will see some higher temperatures... In case it comes you can bring in some serious cooling system... or switch the CPU...
However for now the 65 C is more than perfect...
"The only place where Success comes before Work is in the dictionary." Vidal Sassoon, 1928 - 2012
|
|
|
|
|
It would be, except I have a problem which might be related.
My CPU is performing very poorly on userbenchmark compared to other people's Ryzen 7 4750Gs on the CPU benchmark. It's not anything obvious like slow RAM or throttling by Windows power settings.
So it almost makes me wonder if my BIOS is undervolting my chip or something?
Real programmers use butterflies
|
|
|
|
|
Except the benchmark problem - do you see an actual problem while working on the computer?
Do you know the site 'tom's hardware'? It is a very good place to ask questions about hardware...
"The only place where Success comes before Work is in the dictionary." Vidal Sassoon, 1928 - 2012
|
|
|
|
|
Tom's hardware? I didn't think that site was legitimate. I got malware from there once I think, but it was years ago.
I don't see a problem per se, but the numbers are so out of whack it makes me wonder if there *is* one, you know?
I don't like the idea of hardware anomalies on my system. Smells like trouble to me.
Real programmers use butterflies
|
|
|
|
|
I got no malware from that site, but some good answers...
As for hardware anomalies... Benchmark is not exactly rocket science and if beside it your experience is good I would not bother myself over it... Run your most demanding game and see how it behaves - that is a more solid test than any benchmark software...
"The only place where Success comes before Work is in the dictionary." Vidal Sassoon, 1928 - 2012
|
|
|
|
|
Well, my single core performance *could be better* but that's more about shortening compile times.
With the monster video card I have in it, it eats games.
Real programmers use butterflies
|
|
|
|
|
Do you use -m (or any other definition of parallel compiling)? It may help not to use a single core for all the compilations...
"The only place where Success comes before Work is in the dictionary." Vidal Sassoon, 1928 - 2012
|
|
|
|
|
GCC doesn't seem to be thread aware, but your build system might be. For example, gnu make takes -j to specify how many jobs to run simultaneously. The man page says "If the -j option is given without an argument, make will not limit the number of jobs that can run simultaneously.", which sounds like you'd run into context switch issues if you have a large project with many large modules. Visual Studio can do multi processor compiles tooEnabling Multi-Processor (Parallel) Builds in Visual Studio • Helge Klein
Keep Calm and Carry On
|
|
|
|
|
IIRC -m is for msbuild... If you need on GCC you should use make's -j to do that...
"The only place where Success comes before Work is in the dictionary." Vidal Sassoon, 1928 - 2012
|
|
|
|
|
I like HardwareLuxx too (german forum, but you find english threads as well)...
they were really helpful getting my graphics card and I found several guys to be really competent with AMD topics (however there are a lot of wannabes, as anywhere else in the internet)
M.D.V. 
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
honey the codewitch wrote: My CPU is performing very poorly
Did you try pressing the 'Turbo' button on the front of the computer? 
If you can't laugh at yourself - ask me and I will do it for you.
|
|
|
|
|
Checked the bios version? It can be a bit of a pita the way e.g. asus has organised the bios updates, but I do think it is worth it. Esp. when you upgrade your cpu on the same MB, which is often an option with AMD.
|
|
|
|
|
One of the first things I did after I noticed the issue was update the BIOS. It made no difference.
Real programmers use butterflies
|
|
|
|
|
I have an AMD Ryzen too! 
(AMD Ryzen 7 3700X 8-Core Processor, according to the device manager)
I have no idea how to check temperature or run performance test.. but if you send some link my way I could run the same thing on my machine, as a comparison, if you like?
Caveat, I am using Windows 11, if it makes any difference...
I also have virtual hardware on (for Windows Sandbox!  )
Found it! This thing right? HWMONITOR | Softwares | CPUID (oh this is just the monitoring / temperature thing)
modified 23-Dec-21 5:55am.
|
|
|
|
|
You meant to post it to @code-witch I believe...
"The only place where Success comes before Work is in the dictionary." Vidal Sassoon, 1928 - 2012
|
|
|
|
|
|
Curiously enough, I left my AMD Ryzen laptop running when I was out for a couple of hours and it had rebooted. I saw "DHCPv4 client registered for shutdown notification" in the Event Viewer and googling this suggested it might be an issue with the temperature getting out of hand, but all I had open was some browser windows and Spotify, not even a game or an IDE.
What app do you use to check the stress test? I can try that later today after I log off and get back with my results.
Cheers,
Vikram.
|
|
|
|
|
It's called HWMonitor by CPUID - the same folks that make CPU-Z
HWMONITOR | Softwares | CPUID[^]
Edit: Now I'm hearing that tool doesn't report temps correctly on AMD systems so I'm trying this one: HWInfo64[^]
Real programmers use butterflies
modified 23-Dec-21 2:13am.
|
|
|
|
|
You might do the stress test, a fast reboot and read the temperature in the bios. Is going to be a bit lower, but it should be accurate
M.D.V. 
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
HWMonitor says the temperature is in the 45C to 75C  with just some browsers and Spotify running, plus a remote desktop connection into my office PC. No games of any sort running.
HWInfo has two entries called CPU Termal Trip Limit and CPU HTC Temperature Limit, they are both set to 115C 
The processor is AMD Ryzen 5 3550H, I bought it in Aug 2020.
Cheers,
Vikram.
|
|
|
|
|
Thanks! It seems yours gets up hotter than mine under light load. Mine rests 35C under light load, around 60 on moderate load, to 65 under heavy loads.
That's what's getting to me. This CPU isn't being driven near as hard as it can be, and it's underperforming in my benchmarks. I'm guessing at this point that it's a mobo/bios issue. I'm running the latest BIOS which I upgraded to try to fix this, but it didn't change anything. There might be a setting somewhere.
Real programmers use butterflies
|
|
|
|
|
|
Thanks! I poked at it, and my EDC pegs all red under load, but I guess that's normal.
It hasn't helped the mystery of the underperforming/undertemp issue, but it did show me that the CPU is indeed clocking to full speed under load.
This isn't a big deal - i mean I do wish single core performance was a bit better for long compiles, but mostly this thing is a MONSTER. Even limping along like it is it blows the doors off of a lot of other CPU/APUs.
My main concern is that I may have something misconfigured, and I don't like the idea of undervolting my chip too much, or something like that.
So I'm just trying to figure out the why of this.
Real programmers use butterflies
|
|
|
|
|
Get reminded of an anecdote told by a Computer Science Professor:
In the 1980s, there was an advertisement in a computer magazine that "Our chip is so modern and great that it can complete an infinite loop in 3 minutes."
Reality was that this chip would melt away in 3 minutes while performing the infinite loop computation. So, in essence, the chip completed the infinite loop in 3 minutes.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
