I did some testing on this a while back. I didn't see any benefit, and found I was better off just setting processor affiliation to the even cores for the process I wanted to boost and also upped its priority to force other processes to be scheduled on the odd cores.
As for the 256bit FPU, that happens automatically as far as I know depending on whats running. The FPU on the FX series is weak compared to most Intel chips so don't expect miracles when it is running in 256bit. One way to test this is to use Intel Burn Test as that is hits the FPU.
Some people say that sharing the FPU on a module was a poor design decision. In my opinion as most of the work a PC does is integer based AMD's logic had method to it, and there are other reasons for worse per core performance than recent Intel processors.
Also, shared cache is not always bad; if two threads of the same process are working with the same data you can benefit if the data is already held in the shared cache and does not need to be fetched from memory.
Having said all that I'm keen to hear how the OP gets on :-)
Lolz dude, if you're going to challenge Angel's knowledge on this sort of stuff, you're going to have to get out of bed early and eat a lot of cereal every morning... she is THE resident hardware expert, wait, scratch that, hardware goddess... and until you can demonstrate that you know your shit as well as her, you'll show respect, courtesy and politeness... and what you call "quasi-polite" IS NOT CUTTING IT... and if you don't shape up, you'll get cognitive resonance...
Not sure if you didn't read or if this is just cognitive dissonance on your part so you have an excuse to show off the information you know
If anything flaming, but how is correcting someone flaming?
Lolz, I do read perfectly well thank you...
And before you think I've reacted to your posts because I'm just closing ranks with other mods (which I would do anyway by the way), take a deep breath first and do me a favour: go through the hardware forum and check out all of the knowledge Angel has shared with the forum, and all of the work she put in replying to hundreds of questions... then think again if you really want to call her out in a flaming sense for showing off knowledge... because "showing off" knowledge happens to be a preferred behaviour here from my point of view, not eligible for any sort of flame...
Why is it that there is always resident aggression in those Intel v. AMD threads? Chillax, we'll probably all be using open source RISC designs five years from now anyways lolz... just zip it now because it's not worth it, m'kay... everyone friends, we all play nice, the sun is shining, it's warm outside, we're going for a swim and then we're all hanging on a terrace having an ice-cream with a cold one...
Speaking of which, I'm gonna grab me a piece of cake and a cup of tea...
While I did accuse her of trying to show off, that was not my original intent of my post. My intent was to question why she assumed that the OP didn't know what he was doing.
While I appreciate her contributions to the forums, what she said above was unwarranted as it was presented - directed at the OP. What she could have said was something like - "Hey, just so everyone knows, Asus boards do not have that feature. So you can add that to your list OP."
I do not see what is so hard to comprehend about that and I will not explain myself further. If I was the OP, I would have been offended by her choice of words. Who knows, he may not give a shit, but others, if in his situation, might.
i duno man, i think she was just trying to inform other people who might be thinking of doing the same thing that it wont be possible with an asus board. kinda just throwing out info into the ether, not trying to confront OP. but as OP has not presented any evidence (like his own hardware specs besides having a 8350) this whole thread is kinda suspect anyways.
+1, it might be just another disguised AMD bashing thread and it might be just another typical incident that comes with such threads.
We'll see if the OP comes through with results of his experiments, and then we'll know what kind of thread it really was, but in the mean time, the prediction Angel did on how it would affect performance, seems perfectly logical to me, not that it would even fascinate me to know how it would affect performance, because I really couldn't care less lolz, because I'll worry about performance as soon as all the features that I need start working on Intel like they work on AMD, and I can imagine that for software console users, it's pretty much the other way around, so I think it's just a pointless spitting contest to be honest...
I'm getting close to finishing all the tests, some good, some bad, a lot indifferent. But nonetheless there will be numbers in case someone is curious in the future about doing this. I will be posting a link to the doc with all the tests that I ran and as much information as I can give.
Thank you for the patience and curiosity. And, please remember, this is out of pure curiosity that stemmed from a conversation on teamspeak :)
Just saw this post I've been waiting for someone to this for about a year. I wanted to do this about a year ago and I still haven't been able to raise enough money for an AMD build. I look forward to seeing the results and thanks for sharing.
the performance diffrence per core between amd FX and intel haswell is realy huge.
If you own both a FX8350 and also a 4690K, then its fun to do a Cinebench R15 single threaded render test, and then check how huge the diffrence is per core.
Just to give you an idea how big the diffrence is single threaded.
