(02-16-2013, 03:03 AM)admin89 Wrote: [ -> ]
Lol , I'm just kidding . I bet 20$ the difference is not noticeable
Lol I wouldn't expect it to be too much better
Thanks for all the support people
admin89 Wrote:You didn't read benchmark right ?
i3 3120M / i5 2410M knock out both A10 4600M & A8 4500M in heavily threaded applications .
As I mentioned before , even the low-end i3 370M kill both of them in heavily threaded applications
That's a bad comparison to use since neither the i3 3120M or i3 370M appear in either benchmark not to mention only two applications are used, one of which is synthetic. Still by doing some IPC calculations I can adjust the i5 2410M results to be similar to what I would expect from a 3120M and it does slightly beat the A10-4600M in both tests. Since x264 encoders are pretty much the perfect application for trinity I think it's safe to assume that you are correct and that the i3 3120M will win even in multithreaded applications.
admin89 Wrote:Mobile Trinity 4 cores 4 threads < Mobile Core i3 2 cores 4 threads (Edit : Never mind , I just found out AMD Trinity use something like "Intel Hyper Threading" unlike LLano (this is why Llano outperform Trinity in some benchmark ) . Trinity Dual Core A6 is actually single core , A8 A10 Quad Core is actually dual core
I'm beginning to wonder if anybody reads my posts about bulldozer, zambezi, piledriver, vishera, or trinity. It is not like HT. An amd "module" is not the same as an Intel "core". You need to read up on cluster multi-threading (although most sources don't call it that, they prefer to use the term modules whenever possible).
admin89 Wrote:Desktop - Trinity can trade blow with Desktop 3rd gen i3 and kill i3 in heavily threaded applications . Their mobile CPU is the opposite though
Which suggests that they're still having problems matching Intels energy efficiency (performance per watt). Makes sense since historically this has always been the case since pentium M.
admin89 Wrote:A8 /A10 vs i3 3120M/i5 2410M via Dolphin/PCSX2 ? The difference is something like heaven and hell
I hate to redo this over an over . @OP If you take time and read some benchmark rather than watching some stupid youtube video without "FPS counter" (If they use Frame skipping in the video , they will lose FPS but it can make "not smooth gameplay" become "smooth gameplay")
Note : some i7 in the attachment are low-end ultra low power dual core CPU (which is weak). For example , i7-2637M = i3 2310M
The only thing I'm concerned about is the HD 4000 is slow enough to bottleneck some games and prevent them from reaching fullspeed in my experience. Even at 1x. He's kind of screwed either way.
admin89 Wrote:I even recommend mobile LLano (A6/A8 3xxxM) over mobile Trinity(A6/A8/A10 4xxxM) because LLano can be overclocked and those CPU somehow outperform trinity ? Wth is going on ? AMD . For example , A8 3520M eat A8 4500M alive and it even chop off A10 4600M in Cinebench R10 32bit Multi
Well obviously an OC llano will beat a stock trinity. But that's not a fair comparison. You should OC both if you're going to compare them based on OC performance.
admin89 Wrote:Nope , Overclock a multi-media laptop is not a good idea . Do you want to cook that small form factor motherboard ?
Aren't the A6/A8 also multimedia (multimedia is a valid composite word in English by the way) laptops? Wouldn't your above logic apply to them as well?
admin89 Wrote:I look forward to seeing 3rd gen 2.5GHz beat the heck out of 1st gen 2.9GHz
It should be about 10% faster if I'm not mistaken.
After re-reading some of this thread, I realise that 1/2 core block looks a lot more like half core block than one or two 'core' block.
Sorry to NV for any confusion.
Quote:After re-reading some of this thread, I realise that 1/2 core block looks a lot more like half core block than one or two 'core' block.
What the hell are you talking about? Show me this "1/2 core block" that you speak of.
Quote:Sorry to NV for any confusion.
I'm even more confused now. Before I at least thought I knew what you were trying to say.
I meant module. You were right about what I meant, I just realised that I made it harder to read and understand than was necessary.