AMD760 @ Anandtech!

[crYnOid]

Here

 

[jinsonxu]

Already read, hehe, the IRC channel is fast.

 

[Yza]

yeah.. anyone else notice this line though?

From what we?ve seen though this level of performance won?t be enough to compete with the Pentium 4, luckily AMD can also compete on the basis of price and they also have another ace up their sleeves which they have yet to reveal. -anand

 

[FastEddie]

I don't know how Anand can even make that statement, as the available reviews show a 1.4GHz P4 only about 10% faster on some of the tests when compared to a lower clocked Thnderbird----and the T-Bird was using standard PC133 memory on a KT133 board. I'd like to see that P4 compared to the reference test of the 133MHz fsb 1.2GHz T-Bird with PC2100 memory.

 

[xtreme2k]

i think his SPECs scores are total rubbish
the breakdown of it was really boring

Edit:
also didnt show difference betweeen PC1600/2100

 

[NFS4]

<< i think his SPECs scores are total rubbish >>


Ohh, so Anand just made stuff up?? Yeah right

 

[Bakwetu]

Best review I've seen so far (only read sharky's and Gamers depot so far). I have to disagree with xtreme2k, the specs score was very informative (unless you're implying there is something wrong with them? Explain what in that case.)

 

[jpprod]

Hmm...


<< With the Pentium III on the 840 offering at least a 12% advantage over the Athlon with DDR SDRAM, you can expect the data accesses to come in a very serialized fashion and thus
taking full advantage of the i840's 3.2GB/s of memory bandwidth. >>


Bandwidth on i840 can never peak at over 1.067Gb/s, which is P3's FSB speed. i840 does not take full advantage of 3.2Gb/s bandwidth given by dual channel PC800 DRDRAM, with possible exception of fetches to i840's built-in prefetch cache. Reasons for i840's excellense in 179.art of SPECFP2K must lie somewhere else...

All in all, I'm quite dissapointed with AMD-760's performance. Real-world gains are respectable (such as 10-20% gain in Quake III), but I expected AMD-760/PC2100 Sandra Stream benchmarks to be roughly twice that of a KT133/PC133 system. Perhaps upcoming CAS2-latency PC2100 SDRAM gets closer to this goal.

 

[KDOG]

Seems like the 760 isn't all that....might as well put your money into a faster CPU...

 

[Sephiroth_IX]

W00t.

 

[BurntKooshie]

jpprod Um...why were you expecting double the bandwidth? Theoretically, at the very lowest of levels, yes, it is twice the bandwidth. A case and point as to why we should have expected these results: the G450. Sure its a graphsics card but....

A 64bit DDR bus is theoretically as fast as a 128bit SDR bus (well, in the sense that they both can do the same number of transfers per clock). But, if one looks at the G400, and the G450 scores, the G450 scores are lower. Same chip (well, the G450 has more stuff on it, but its not like they took anything out), same clock speeds for core and mem, yet the G450 scores lower. This means that DDR is less than twice as fast as SDR.

So when the clocks are the same (133mhz SDR v. 133mhz DDR), the difference should not be in favor of DDR two fold, rather, somewhere lower.

 

[Thorn]

<< In comparison to the upcoming Pentium 4, even with DDR SDRAM and the 133MHz FSB it may be difficult for a 1.2GHz Athlon to compete with a 1.5GHz Pentium 4 in many benchmarks. AMD may be forced to release a higher clocked Athlon in order to remain competitive performance-wise, however they will almost definitely hold the price to performance ratio crown throughout the rest of the year. >>


Yza, which page had the statement you quoted? I couldn't find it. The above is the closest thing I could find, but it simply states that the 1.2GHz Athlon may not be quite as fast as the 1.5GHz P4 in some situations. On a clock for clock basis however the Athlon will still be faster.

 

[jpprod]

jpprod Um...why were you expecting double the bandwidth?

Of course not double real-world bandwidth, but I was expecting the nearly optimal situation for RAM bandwidth (namely Sandra's stream benchmarks) show improvement in the range of 70-80%. Instead we got 40-50%. Respectable certainly, but not earth-shattering.

Then again, you should all read Tom's AMD-760/DDR review. It points out that no other new&amp;upcoming memory type (EDO vs. SDRAM, PC66 vs. PC100 etc.) has ever made this much difference in real-world performance as PC133 vs. PC2100 does.

 

[XMan]

Great point in your last post, Jpprod!

I think the thing everyone is overlooking is that Mustang and Palomino should be released shortly. I'm not jumping into the pool until all the toys are out . . . probably around January?

 

[Yza]

Sorry I cant find the exact spot right now I'm off to my first day on the job.. I'll search back later, and post.

 

[Zorba]

Thorn: Yza's quote was from this page

 

[Shuxclams]

<< Intel?s current fastest setup is the i840 with the Pentium III 1GHz, and the 1GHz Athlon on the AMD 760 has no
problem trampling all over that solution. A combination of the new 266MHz FSB and its PC2100 DDR SDRAM provide
for this 25% performance lead. >>



Clock for clock the fastest soluotion out as of today is Athlon+DDR.




SHUX

 

[SSP]

UnrealTournament is showing us an 8% improvement like MDK2, and we also see the dethroning of the Pentium III as the performance king on a clock for clock basis.

What I was waiting to hear. All the other synthetic benchmarks mean sh!t to me.

 

[HaVoC]

I'm surprised at the lukewarm response here. I'm very satisfied by the memory performance of AMD 760 DDR. Gains of 10% to 20% performance real world are very good for memory.

 

[Soccerman]

hmm.. I'm not surprised. the benches look about right.

I'm getting more curious every time I see an Anandtech review. what kind of BIOS settings do they use (there are alot that could affect performance).

as for the P3 being faster in games, well that is most likely because of better optimization (ala Quake 3!) for SSE then 3DNow!