TR Forums

About me just giving this up...

I was pretty happy with it all until this S3 realization.
In the end, that I keep or dump this, I just want to understand the problem.

My latest finding: In the Bios I have an "OS" or "Bios" option under S3

With "OS" is selected, Turbo will always be enabled out of S3 (and in concert APM)
When I switch the option to Bios, turbo is NOT enabled coming out of S3... but APM is.

My test is simple.
a) Run prime95 torture test while AIDA graph the CPU clock for a few minutes untill temp settles
b) Put the computer to sleep, and wake it up
c) check the clock graph

In my case the graph is rock solid at 4.2ghz before sleep mode, and bounce betwen 4.2 and 2.9ghz after a sleep cycle.

If I load AMD Overdrive (S3 Bios), turbo is shown disabled after S3. Now, if enable it and disable it... Poof, the clock graph is not again rock solid at 4.2ghz

The throttling I experience is 100% confirmed to be APM related.

So any motherboard that would allow me to disable APM and keep it disable out of S3 would solve it for me.

But that doesn't explain how Turbo throttle even at stock... Is APM actually managed by the motherboard?
If APM power throttling is a motherboard feature, then AMD is not at fault, beside letting OEM built shoddy motherboard.

After ranting about this, I was told "Why dont you forget about sleep mode and just hybernate?"
I didn't think about it and yes,this solves the APM issue not being restored, I'm not excited about it but I will set this up.

Another rant. My Q6600 come out of S3 sleep under 5 second (mouse move to be back in my apps), the FX8320 about 11 seconds.
Seem like we went backward in the past 7 years in that respect... Unexpected as S3 is a core feature for mobile computing.

AMD doesn't "let" OEMs build shoddy motherboards, unless you're seeing this motherboard on some recommended list on AMD's site. (Did you? Genuinely curious.)

AMD publishes the specs for their sockets and chipsets. OEMs implement to those specs, to the best of their abilities. MSI is (within a factor of 2 by total revenue) roughly the same size as AMD, but most of their revenue comes from making motherboards for Intel CPUs, not AMD. I doubt AMD is in much of a position to dictate how MSI behaves.

As far as the hibernate times are concerned... are you even comparing apples to apples here? Time to resume from hibernate is almost entirely determined by the amount of RAM you have and the speed of the drive you've put the hibernate file on; it has very little to do with the CPU. Unless you've got the same amount of RAM and a comparable speed hibernate drive across the two systems, direct comparisons (without correcting for those factors) of resume-from-hibernate times between your Intel and AMD system are meaningless.

.

sschaem wrote:

AMD is not at fault, beside letting OEM built shoddy motherboard.

10-74, sschaem. You could be 10-56 or 10-96, it would take a 10-81 to be sure, but either way: PEBCAK. I'll 10-9 what others have said: shoddy mobo's are not AMD's responsibility; if it's shoddy only MSI can be blamed. Over.

I can give you plenty of things to criticize AMD for, but a shoddy MSI motherboard is not one of them.

.

For anyone playing around with their FX on a 'feature limited' (to say it nicely) motherboard.

http://www.xtremesystems.org/forums/sho ... w-Versions

APM=0 works for me. Also it seem that you can keep turbo and disable APM ! So you can keep you TDP low, yet overclock, and also have fast single threaded performance !

So I'm all set. I now have a little icon on my desktop to disable APM when I need to.

In reality its not a huge issue because I decide to keep the CPU at 4.2ghz, and even loading the CPU 100% with code like Cinebench doesn't push the power limit that bad.
So even if APM is turn back on after sleep, in my case I lose about 5% performance at most in real apps.

Literally Case closed (and thumb screw tighten)

Running AIDA64 overnight, and will re-image clean tomorrow... This has been an interested process.

Those FX processors have a lot of potential, it to bad AMD decided not to continue tweaking the architecture for desktop.

clone wrote:

and what about the multiple bios updates that every motherboard ever sold has gotten.... are those AMD and Intel's fault as well?

Actually in a number of cases it IS AMD and Intel's fault.

.