• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Processor architecture and databases performance

X

XcomCheetah

Junior Member
I m not sure if this is the rite forum to put this question but i m doing a term paper in Database course and i m thinking of doing a performance analysis of databases on different architecture... i know that the architecture are not designed with respect to database but database do constitute a huge portion and different architecture specially server side architecture like Intel's EPIC and Sun's Sparc and IBM's Power processor must have some designs relating to increasing the performance of databases.... can any of u guide me to some resources where i can find more information abt this... also i was thinking of these new concept of Hyperthreading and SMT and dual core processor and if i remember there was some research going on putting a secondary chip to offload the burden of I/O from main memory.. how these techniques helps databases..
Any kind of quidance will be appreciated..

Thanks.
ps. i have only read x86 architecture so i m pretty blank on rest of the architectures but i m more than willing to get acquainted with them.🙂
 
I suspect the biggest cpu issues wrt databases will be on chip cache, and, for really big DB, a 64 bit address space.
 
I'm not a sysadmin or anything. But I'd guess with databases the pipes would be more important than the switches. Generally database calculations are simple calculations that must be performed repeatedly over a vast array of data. So large caches, lots of ram w/ fast interfaces, and drive arrays that would make any geek swoon would be more important than raw CPU speed.

Kind of similar to videocards and console gaming systems really. In the last couple generations of vidcards (GF2-4, Radeon) the main limitation of speed isn't the GPU, but the RAM. Tests have shown that OCing a GPU from 220-275MHz doesn't yeild many improvements, yet OCing the RAM by the same amount yeilds fantastic results.

Why is this relevant? The GPU performs the same calculations over and over on a vast array of data. Just like a database CPU. (but a bit more specialized).
 
and wot abt the RAM latency... i mean will high speed RAM like RDRAM will help or low speed and low latency DDR ram is good... and secondly wot abt the hyperthreading technology..?? any improvement from it..?? and wot abt the benchmarks... are there any databases performance related benchmarks.??

Thanks for all the help..
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top