(To Tom) P4 and Bapco benchmark

[Guest]

I was really surprised when I saw P4 scoring better result then Athlon in a web-based benchmark, because I thought (as I still do) the mayor bottleneck in todays internet application is the internet connection itself, not the CPU or the RAM.
So, why should a CPU which ony shine in bandwitdh intensive applications shine also in applications handling as low as 56 Kbps??
I suppose there's something wrong in the test process.
I think the possible gotchas are:
1. connection speed: Tom run the benchmark over a LAN. I suppose 100 Mpbs. If there's not some kind of throughput control, I think it's not a realistic setup for most of the users.
2. streaming content: how big (= how much bandwidth) were the streaming content? In example, were the QuickTime movies suitable for a 56 Kbps connection?

Kind regards,
Corrado

 

[Grizely1]

"Internet Applications", not "Internet". IE, etc.

 

[AmdMELTDOWN]

you have a point, tmg should do test on that type on connection speed but I'm pretty sure that the majority of maket who will benefit from intel internet optimized cpu will be businesses who have high speed connection greater than 100Mpbs. I'm also hoping that maybe the folks at ID can patch their quake engine to take advantage of the P4.

 

[Guest]

good point. we definitely need more detailed knowledge on what the BAPCo web benchmark is really about.

i remember Tom mentioned it is a client-side test, a client connected to a <b>dedicated</b> server. and i guess the test was done with broadband connection. it seems to be more like an intranet than the Internet.

-----------------------------
Some are ignorantly happy,
While some, happily ignorant.