0
I have an AMD FX-6300 CPU, non-overclocked.
I have all power saving features such as C1E turned off in the BIOS. I have also disabled turbo-boost.
At stock settings, this CPU should run at 3500Mhz(3.5Ghz). However, using lscpu in Ubuntu shows that the CPU is running at 3511.588Mhz.
I should also mention that HWmonitor on Windows 7 reports similar results.
Does my CPU just have an awful clock crystal? What's going on here?
You'reAGitForNotUsingGit
Posted 2016-03-29T14:32:31.773
Reputation: 614
1clocks rarely every operate at a frequency that ends in
.0if you read it with any real precision. Most apps round to the Tenths position, or even to the ones position, but it is the app that is reducing the precision. any reading that gives you a nice round number is more suspect than one rounded to the thousandths or beyond as you show. – Frank Thomas – 2016-03-29T15:01:30.433@FrankThomas - But on a different machine using a different CPU, lscpu shows the clean max freq of 2800Mhz. – You'reAGitForNotUsingGit – 2016-03-29T15:03:34.100
lscpu has hardware-aware code in its base, so it will vary based on the specific hardware. the Mint system I'm using now is reporting 3390.417 for a 3.4GHz box. Family 6/Model 60/ Stepping 3. – Frank Thomas – 2016-03-29T15:06:15.240
@FrankThomas - Meaning that lscpu is lying to me, and my other machine is not cleanly running at 2800Mhz? – You'reAGitForNotUsingGit – 2016-03-29T15:11:03.250
Well, first I would say that it is essentially impossible that any software can tell you the "truth" if you want to put it in a boolean context. In analog reality, almost nothing comes out perfectly to 0 or 1. In fact, as a Programmer, when I write code to do something like (sqrt(2) * 33.31), and run that code twice, the 9-decimal floating point response would be different. From a digital perspective they shouldn't be, but these machines are analog, with polarized streams of electrons flowing through them like water in a pipe. water doesn't come out of a pipe at a precise constant rate. – Frank Thomas – 2016-03-29T15:25:44.040
@FrankThomas - Very interesting, because I was having problems with some of my Java code overshooting a value by a minuscule value (+.000000001), which led to an if-statement failing. Now I think I know what's going on! – You'reAGitForNotUsingGit – 2016-03-29T15:33:02.363
@AndroidDev - Software cannot "lie" it can only do precisely what it was programmed to do. Unless you start comparing the same Apple running on different hardware is being reported differently your statement isn't accurate. The clocking being report is accurate. – Ramhound – 2016-03-29T16:04:24.063
@Ramhound - But in that case we're right back to my original question: Why is the FX6300 not running at clean maximum frequency, while my other machine's Athlon CPU does? – You'reAGitForNotUsingGit – 2016-03-29T16:08:24.663
It is a clean frequency, unless you are getting different readings each time, then it is clean. The obvious answer, as the reason for the differences in readings, is the hardware code works differently. You can't compare Apples to Oranges, the hardware code, likely is not the same. – Ramhound – 2016-03-29T16:11:49.607