RadioHead84

Platinum Member
Jan 8, 2004
2,166
0
0
My friend is thinking about builing a new pc...

around 2grand to spend..

He just told me he was going to do a dual opertron 1.8gzh...

Is this wise?

Will it be as fast as a 3500? Will it out perform? How much will it cost? Will it run Windows 64bit?

I know nothing about duals!
 

Vette73

Lifer
Jul 5, 2000
21,503
9
0
WOW!!! Wish I could get a Dual Opteron system.

But yes it will be as fast or faster then a 3500+ in programs that are SMP, but slower on games as they can only, for the most part, see 1 cpu.
And yes it can run 64bit OS. It can run Linux right now and the Beta WinXP-64
 

CraigRT

Lifer
Jun 16, 2000
31,440
5
0
Originally posted by: RadioHead84
So what programs are SMP...will they be?

some programs have native SMP support, usually development apps.
the rest of the programs are usually assignable. such as distributed computing apps. you can make them run on both CPU's at once, which is my entire reasing for liking dual so much... so much flexibility.
my next system will be a dual Opteron no question.
 

Yanagi

Golden Member
Jun 8, 2004
1,678
0
0
the added extra cost and that most games and average joe apps doesnt support SMP.

If you dont want to do dual now, wait for dual core CPUs- The same thing but 2 cpus on one chip.
 

aweber1nj

Junior Member
Aug 19, 2004
4
0
0
Check this past week's PCWeek. They did a comparison test of dual Xeon vs. dual Opteron Wkstations. Apparently the Opteron is excellent when you're single-threading (anyone, anyone? Buhler?), but as you increase the number of procs running in your OS the Xeon blows it away.

Would splurge for the new EM64 flavor of Xeon if you can fit it in the budget, but the test was the 32bit Xeon vs. the Opteron (again, I think it was running 32bit Win2K).
 

carlosd

Senior member
Aug 3, 2004
782
0
0
Originally posted by: aweber1nj
Check this past week's PCWeek. They did a comparison test of dual Xeon vs. dual Opteron Wkstations. Apparently the Opteron is excellent when you're single-threading (anyone, anyone? Buhler?), but as you increase the number of procs running in your OS the Xeon blows it away.

Would splurge for the new EM64 flavor of Xeon if you can fit it in the budget, but the test was the 32bit Xeon vs. the Opteron (again, I think it was running 32bit Win2K).

No Xeons are not faster anyway.
Actually multiprocessing opterons are best performers in world, even over the G5 plataform.

The reasons:

Bottleneck problems are solved with Hypertransport and the built-in memory controller. The Opteron architecture easily scales to eight CPUs.

Traditionally, communication between the hard drives, AGP/PCI cards, memory and CPU have been slowed by having to pass through various controller and Bridge Chips on the motherboard. With Hypertransport, AMD has allowed components to talk to each directly.

The on-CPU memory controller lowers latencies because it gives direct access to RAM. The memory controller also clocks in at CPU speed. Multi-processor systems have traditionally been hampered by memory latencies - after all, you have to feed all those processors information. This is where the competition can't keep up.
 

aweber1nj

Junior Member
Aug 19, 2004
4
0
0
No flames, please.

Just relaying what I read about an independent company (who certainly takes advertizing dollars from them all) who has the means to purchase a bunch of OOTB systems and run industry standard benchmarks on them wrote.

If you read the article, I think it surprised them as well. But the bottom line is for multi-processing the Xeon out-scaled the Opteron.

BTW: My-bad, it was InfoWorld that did the comparison; pg. 24 of the Aug 16 issue. Also at: InfoWorld Article
 

CraigRT

Lifer
Jun 16, 2000
31,440
5
0
Originally posted by: RadioHead84
so why isnt everyonedoing it if its so great.

I am personally getting the 3500 set up

availability of hardware, and cost of hardware perhaps?
I had a Dual MP1800 for a while (only a month or 2 ago) but it was too slow for gaming and such... so I don't have it anymore, but I will go back to dual when I can afford a dual Opteron.
 

justly

Banned
Jul 25, 2003
493
0
0
Originally posted by: aweber1nj
No flames, please.

Just relaying what I read about an independent company (who certainly takes advertizing dollars from them all) who has the means to purchase a bunch of OOTB systems and run industry standard benchmarks on them wrote.

If you read the article, I think it surprised them as well. But the bottom line is for multi-processing the Xeon out-scaled the Opteron.

BTW: My-bad, it was InfoWorld that did the comparison; pg. 24 of the Aug 16 issue. Also at: InfoWorld Article

I suspected it was either a copy of the InfoWorld article or you had the source wrong. This article was discussed in another thread a week or two ago. The author of the article would not (or rather could not) answer some legitimate questions regarding his findings. While I have no doubt that he reported his findings correctly there is serious doubt about how well the test suite represents real life scenarios.

The real question you should be asking is how much multitasking is required for a duel P4 to overtake a duel Opteron. It doesn?t tell that does it? It also doesn?t tell if this multitasking is on something the P4 is already better at (like encoding). Nor does it tell what ratio of these multitasking apps are streaming or branch intensive, light or demanding on bandwidth, or even if the multitasking load the P4 excels at is considered normal or highly exaggerated.
 

carlosd

Senior member
Aug 3, 2004
782
0
0
Originally posted by: aweber1nj
No flames, please.

Just relaying what I read about an independent company (who certainly takes advertizing dollars from them all) who has the means to purchase a bunch of OOTB systems and run industry standard benchmarks on them wrote.

If you read the article, I think it surprised them as well. But the bottom line is for multi-processing the Xeon out-scaled the Opteron.

BTW: My-bad, it was InfoWorld that did the comparison; pg. 24 of the Aug 16 issue. Also at: InfoWorld Article

The article is only informative, and doesn't prove anything, since they don't show us the benchmarks, or the applications they have used, they can't prove nothing. It seems like an intel- paid page, it seems like intel is desperated because they are about to lose the server-workstation crown, too.