New Posts  All Forums:

Posts by cssorkinman

I meant H/T effiency vs Ryzen's SMT in this particular scenario.Cpuz multithread, pov ray 3.7 and hwbot prime are a few instances where the FX has a chance against the 4790k.
Are there any added features in the new bios vs 1.1 that you can see?
So in the benchmark examples of core comparisons I posted earlier - P+L = about 70 % of P+P yes?Should fire up the 4790K and compare I guess.
That's not too far from how I pictured it . The FX platform's behavior influenced how I thought affinity should work, I had observed the differences between it and my i 7's but wrongly interpreted what I was seeing. At any rate , thanks for the information. Always something to feed my curiosity it seems.
Possible the difference I was seeing was due to that or the fact that the cores were saturated on the FX and weren't on the I 7 and I mis-interpreted what I was seeing - appreciate your time Blameless.
Possible I was seeing windows scheduling workload intelligently to avoid inefficiencies that 2 cores with H/T would express vs 4 cores ?
At some point in all my messing around with the I 7's I have - the results I were seeing led me to assume they were set up as I described. I remember it behaving much differently than the FX - Vishera was quite obvious and easy to demonstrate .
Don't I 7's lay out the core affinity in task manager as such - first half = physical second half= logical?
Currently the Titanium's highest divider allows for 3200 mhz and it has no base clock adjustments - other boards can.Sure seems like selecting adjacent cores causes some kind of contention - it would appear that window's scheduling could make a HUGE difference in certain circumstances.Edit: Core affinity Even numbers= physical , odd = logical
Not sure what to make of this other than I need to do more testing...lol
New Posts  All Forums: