Slashdot Mirror
Intel Core 2 Duo Vs. AMD AM2
ThinSkin writes, "ExtremeTech has an extensive performance roundup across the entire line of Intel Core 2 Duo and AMD AM2 CPUs, from the cheap to the ultra-high end. Both companies bring five processors to the table, ranging from $152 to $1,075, with the mid-range CPUs boasting the best in price/performance. From the article: 'It's clear that Intel's Core 2 Duo lineup offers superior performance across the product line when compared with AMD's Athlon 64. In some applications, even a lower-cost Core 2 Duo can outperform some of the higher-end Athlon 64s.'" The ExtremeTech article is spread over 10 ad-laden pages. You can read it all on the printer-friendly page, but you'll miss out on the pretty graphs.
http://libtomcrypt.com/ltc113.html
crypto work done in 64-bit mode on the Core 2.
Tom
Someday, I'll have a real sig.
Clock speed won't change the cycle counts. I'm not counting wall time here. So you can expect that if AES takes 247 cycles @1.83GHz then it will probably take 247 cycles @2.93GHz.
Tom
Someday, I'll have a real sig.
Looks to me like AM2 starts a little lower than $152.
Every AM2 processor is dual-core, or "X2."
I'm not going hunting for the links, but all power consumption comparisons I've seen show intel in the lead over AMD now in terms of power consumption vs. performance.
I'm not sure about dollar for dollar any more, AMD stuff is going cheap now because they've lost the lead, especially if you don't mind relatively poor performance.
dB for db? since when did processors make noise? If you're talking about their respective heat output for equivalent performance, again it seems intel are now ahead. The core 2 runs cool enough that he cpu fan can be easily replaced with a fanless (silent) alternative anyway.
At the low end, the E6300 at $190 beat the $187 AMD 4200+ in all tests, and also beat the $253 4600+ in 3 out of 4 (with the 4th test extremely close).
At the midrange, the $360 E6600 beats even the $825 FX-62 in all 4 tests. That is bad, bad, bad for AMD.
At the high end, AMD simply has no answer to the $559 E6700 or the $1075 X6800.
Granted, none of their graphs shows the ScienceMark. But overall the results seem pretty one-sided to me. I'm surprised AMD hasn't dropped prices more.
Thats nice. But most users use non-scientific applications that will see very little benefit from 64-bit ops. The extra register space is really not a big deal with a superscalar out-of-order where register renaming and high L1 bandwidth keeps the instructions moving. Stack spills/fills nearly always hit in the L1 and performance nowadays is becoming dominated by cache misses. 64-bit is certainly necessary to move beyond 4GB of address space (with segment tricks), but most users will see no benefit from 64-bit ALU operations. In fact, 32-bit datapaths can be clocked faster than 64-bit datapaths.
with DX 10 "regulating" what features can go into the cards, MS wins.
Uh, DirectX is free. Writing DirectX problems with the free DirectX SDK is also, you guessed it, free.
Microsoft doesn't profit directly from DirectX. Instead, by making Windows a better platform for game development they, shock, get more game developers on Windows.
Also note that Microsoft doesn't decide what features can and can't go into a DirectX 10 card - it sets a minimum featureset for cards that want the sticker. How horrible that a card being marketed as supporting DirectX 10 has to support DirectX 10 functions. (Remember that DirectX emulates hardware functions your videocards lack, allowing games written for it to transcend specific videocards. If the videocard doesn't support any advanced texture, lighting, and whatever else features, you really have a DirectX 10 complaint CPU.)
DATABASE WOW WOW
It's an ALU benchmark nothing more. The goal is to see if the ALU has improved or not.
The code all fits in either the L1 or L2 (the test program is less than a 1MB) so it doesn't matter that I used a low end Conroe. The 1.83Ghz conroe has the SAME CORE as the expensive 2.9GHz conroe...
Tom
Someday, I'll have a real sig.