Is there any way to put a filter on this thread so I don't have to read about Sledge's 4.6GHz proc that still has head room. It is deflating my geek-ego.....I bet that would be a sic folder though. /cough /cough
Tried setting the affinity to 1 core and tested all 12 cores. Results were 8.907 - 8.968 never dropped below the 8.8 mark.
On my quad core, SuperPi is faster when I set the affinity on one of the cores. But you have virtual cores and physical cores due to hyperthreading. Affinity should not be on one of the virtual cores. Hyperthreading is causing the difference I guess.
@Sledgehammer70 said:
Bringing this thread back to life!!! We seem to have bypassed a few generations of AMD and Intel CPU's.
So... Not so new Intel i7 4790 running 4.680Ghz or 4.7Ghz.. not sure why it is jumping around between programs.
What's obviously happening is that one program calculates as a whole number multiplier but a fraction/decimal bus speed while the other does just the opposite. So it's not very likely that they would ever match, but it could happen by chance I suppose. "Why?" is still the question here, like you asked.
What would be the most reliable method to gauge your processor speed within an OS? I mean you may find a reliable calculation from within the bios, however that may all change once you enter an OS. Me, I just call it 100x47.
@Sledgehammer70 said:
Bringing this thread back to life!!! We seem to have bypassed a few generations of AMD and Intel CPU's.
So... Not so new Intel i7 4790 running 4.680Ghz or 4.7Ghz.. not sure why it is jumping around between programs.
I might be a few months behind,
coming out of nowhere here....
But just testing my new i7 5820k x99 system.
Overclock at 4.2Ghz only pulls 10.218 seconds.
That seems off for a new gen 6 core to NOT pull a faster score than my old i7 920.
Comments
The cool thing is this CPU still has head room to OC. I am running 4.6GHz as a stock OC and the power draw isn't that crazy.
On my quad core, SuperPi is faster when I set the affinity on one of the cores. But you have virtual cores and physical cores due to hyperthreading. Affinity should not be on one of the virtual cores. Hyperthreading is causing the difference I guess.
It might be time to step things up though, some Dice or POSSIBLY some LN2
There's an Evga P55 FTW board in my closet just collecting dust and it might have to be sacrificed for the cause!
If you go Dice or LN2 you should spank my scores on water... but to make things fun i ran a bit better score at 4.74Ghz
Fixed
Bringing this thread back to life!!! We seem to have bypassed a few generations of AMD and Intel CPU's.
So... Not so new Intel i7 4790 running 4.680Ghz or 4.7Ghz.. not sure why it is jumping around between programs.
I'm pretty sure several generations were passed for Super Pi due to it being an outdated "benchmark"...if you can even call it that now
Side note - 4 year old bump for Sledge
My 3930k does about 10 seconds but I notice only a single core is used. Nice necro.
What's obviously happening is that one program calculates as a whole number multiplier but a fraction/decimal bus speed while the other does just the opposite. So it's not very likely that they would ever match, but it could happen by chance I suppose. "Why?" is still the question here, like you asked.
What would be the most reliable method to gauge your processor speed within an OS? I mean you may find a reliable calculation from within the bios, however that may all change once you enter an OS. Me, I just call it 100x47.
I might be a few months behind,
coming out of nowhere here....
But just testing my new i7 5820k x99 system.
Overclock at 4.2Ghz only pulls 10.218 seconds.
That seems off for a new gen 6 core to NOT pull a faster score than my old i7 920.
I blame it on the program
If the clockspeed on your i7 920 was higher, this would explain it.