Download the Tom's Hardware App from the App Store
The reference for current tech news
Yes No

Conclusion

by

The initial comparisons between the new quad core Phenom and the Intel processors weren’t favorable for AMD, as Phenom cannot compete with Intel’s top of the line Core 2 offerings. Especially out of reach for AMD is the 45 nm Penryn generation, which was presented some weeks ago as the new Core 2 Extreme QX9770, and which will launch into the mainstream in January. Once again, Intel seems to be almost one manufacturing cycle ahead, as it is about to switch from 65 nm to 45 nm. Meanwhile, AMD is still struggling at 65 nm. Consequently, its strategy was adjusted to attack the mainstream.

Despite all the delays and Phenom’s L3 TLD bug, there is nothing wrong with Barcelona from the standpoint of how it was designed. We found noticeable performance advantages in every benchmark, which proves that Phenom is indeed faster than the Athlon 64 X2. In order to get a true core to core comparison, we ran both an Athlon 64 X2 and the Phenom using only a single processing core. This way, we were able to find evidence for AMD’s claims of 25% better performance on Phenom when compared to Athlon 64 X2. While the single core comparison didn’t show that much of an increase, we should consider that you’ll be using three more cores in productive environments.

Some of you may now point to the performance increase Intel was able to achieve when it went from the Pentium D to the Core 2 Duo. Looking at the performance gains and the performance per watt ratio, AMD certainly didn’t make such a leap, but Phenom doesn’t represent such a major generation change either. Let’s hope that AMD can finally fix the remaining bugs with Phenom and get the manufacturing to a level that allows Phenom to run cool and quicker, and to enable the company to be profitable in its core business. Until then, there is no option but to leer at the upgrade market.

Share:
5
Comments
Read more
X
Submit

Comments
Read the comments on the forums
dloneranger 19/12/2007 23:32
Hide
-0+

Interesting, but by only using one core it doesn't give a decent real world test
I'd have like to have seen a real 'upgrade' comparison, eg divx encoding on an x2 vs phenom to give me some idea if it's worth the money

leexgx 20/12/2007 04:05
Hide
-0+

not realy the Point of this test was to show how real world apps perform in Single threded apps

was hopeing anandtech did this but thay used Faster X2 amd chips so there was No point at all included them results unless thay had the same clocked chips as well

and thay have all ready done (or other sites have any way) done divx tests but like i sad before thay needed to inclued same clocked X2 chips as well

untill X4 3ghz cpus are out or can run at it 2.6ghz X4 performs like an 2.7-2.9ghz X2 cpu (apart from when more then 2 cores are used then it smoke all X2 chips no mater what clock speed thay run at, Divix for e.g.)

spuddyt 20/12/2007 12:02
Hide
-0+

DAAMIT I'm now in limbo, wanting to upgrade, but not really sure if i should..... (actually i'm pretty sure that I shouldn't)

wild9 21/12/2007 02:15
Hide
-0+

I still maintain these are better products than Intel's C2Q. A real-world testing using heavy multi-tasking would show the benefits of a monolithic design and in a cluster environment this would simply win hands down, since inter-core bandwidth is essential. Intel's offerings concentrate on nstructions per clock - not bandwidth.

8 of these cores when used in a server would blow away Intel; the memory latency and HTT 2.0/3.0 interconnect would be leaps ahead. Intel cannot fit monolithic cores onto a single die as the design is flawed.

When are we gonna see 128-bit SIMD instructions to really show off what this can do?

splodger 08/01/2008 12:29
Hide
-0+

I urge you all to check out this review http://www.lostcircuits.com/cpu/amd_phenom/ on lost circuits and direct you to the multithreading benchmarks F.E.A.R and especially UT3.

All isn't lost for AMD

Best offers

Newsletters


OK