Calm Down

[Big_Spark]

BAS!! how the eck did you get them outta my portfolio!!!???

lol, na i said computer- as in domestic-and at 3k i would hopw it woud be good

ss

So what spec did you get for 3K?

 

[ban-all-sheds]

EXCUSE me Mr Masona Sir, please don't swear in my direction...Ikea indeed...watch it son, you'll curdle my cocoa

MFI then....

 

[masona]

MFI then....

Can't be, MFI stand for " Made For Idiot "

 

[Big_Spark]

Try purpose made by professional Carpenter.

 

[masona]

 

[supersparks]

duel 3.6Ghz amd athlon- 180GB hard drive
512 MB DDR
GF 4 128(graphics)
and 2 dvd players.
the screen was upgraded to flat and is 28 inch flat screen
came with a printer/scanner - combined unit- when they were new
a and a TV card, few other bits and bobs but the screen was about a grand of that on its own

 

[Big_Spark]

duel 3.6Ghz amd athlon- 180GB hard drive

That can't be right as AMD do not make a 3.6GHz CPU.

The fastest are the XP3200 Barton's, and they actually only run at about 2.7GHz true speed.

You sure they are not MP2600 AMD's? If they are not the MP version then having them in a dual setup is pointless as they are not MP enabled CPU's, like Xeon's, Opterons etc.

 

[supersparks]

ill take a look at the specs again- 3.6ghz...amd may be wrong...im sure the bloke said duel...but i did look a lot of computers that night....

 

[plugwash]

at least one major pc magazine tried a pair of athlon XP's in a dual board and they worked perfectly giving the exact same score as the equivilent MP chips

furthermore they tried thunderbird athlons in there and they also worked perfectly as a pair

 

[Big_Spark]

at least one major pc magazine tried a pair of athlon XP's in a dual board and they worked perfectly giving the exact same score as the equivilent MP chips

furthermore they tried thunderbird athlons in there and they also worked perfectly as a pair

It is true that some XP chips will work as MP's do, Windows XP or W2K Pro helps this along as they are true MP enabled OS's. My P4C's show as dual processors, even though it is a single for this very reason.

However, if you took some true multi-tasking software and ran it at the same time on a machine with a pair of MP chips in, and another with a pair of XP chips in, the MP pair would trounce the XP's.

I do this all the time with SETI, and you will not get a better program to stress test any system.

However when it comes to MP enabled chips, Intel still rules the roost.

 

[plugwash]

i can't remember exactly what benchmarks they were using but they were giving far better results than they did on a single processor of that speed

and the results from the xp's and the mp's were identical indeed the motherboard couldn't even tell the difference! and i'm pretty sure seti was on thier list of tests

i think the xp and the mp are essentially the same it's just the xp is not tested and gauranteed to work in multiprocessor configurations

btw the reason your p4 shows as two cpus is because of hyperthreading which is omething totally unrelated and something no amd chip has ever had

ok i found the peoce in the hands on hardware section of personal computer world febuary 2002 here are the figures (they used a tyan tiger mp board to test on)

sysmark 2001 internet content creation
dual mp 1800+ 250
dual xp 1800+ 250
dual thunderbird 1.2ghz 195
single xp 1800+ 178
dual p3 933 156
single athlon 1.2ghz 131

here you can clearly see the xp and mp pairs perform identically and much higher than the single xp the dual thunderburd isn't putting in a bad show either far better than the single thunderbird

sysmark 2001 office productivity (in sysmark tests higher is better)

single xp 1800+ 180
dual mp 1800+ 179
dual xp 1800+ 179
dual thunderbird 1.2ghz 155
single athlon 1.2ghz 144
dual p3 933 122

the dual cpu setups didn't do as well here probablly because of disk io but the XPs and MPs still performed identically to each other

seti@home (time to complete 1% of a particular work unit they use to test)

dual mp 1800+ 3:17
dual xp 1800+ 3:17
dual thunderbird 1.2ghz 3:50
dual p3 933 4:51
single xp 1800+ 5:55
single athlon 1.2ghz 9:05

here it's even more pronouced than with the first test the difference that dual cpus are making and once again the XPs are performing identically to the MPs

 

[ban-all-sheds]

However when it comes to MP enabled chips, Intel still rules the roost.

Ahem.

I probably shouldn't use this forum to plug my employers products, so no-names, no packdrill, but whilst what you say may well be true for the x86 architecture, there are other CPU families, and some of them scale a great deal better than Intel's or AMD's

 

[Big_Spark]

seti@home (time to complete 1% of a particular work unit they use to test)

dual mp 1800+ 3:17
dual xp 1800+ 3:17
dual thunderbird 1.2ghz 3:50
dual p3 933 4:51
single xp 1800+ 5:55
single athlon 1.2ghz 9:05

here it's even more pronouced than with the first test the difference that dual cpus are making and once again the XPs are performing identically to the MPs

They probably had someone enable the MP part of the XP CPU to such parity in performance, it is easy to do if you know which bridges to short out. I have a pair of Unlocked XP's that will happily run with the multiplier as high as 17..but damn to they run hot.

Those figures for the 1% of a SETI unit are absolute bull, even if they used the GUI version, as I'm sure they did.

My firewall machine, which runs SETI as well, is a dual P3 850 Tualatin machine, this completes a unit in a little over 7hr 22 mins ( two units in reality as it runs two together..one per cpu) Now if that was a single CPU machine of the same spec, it would be slightly faster at around 7hr 05 mins.

If I ran the GUI on this machine, then it would take about 11 hours per work unit, so you can see that the figures they quote are complete bull, and thus I would question all those figures as the integrity of their tests are questionable.

Regarding the P4C's, I am aware that it is Hyperthreading that allows an MP enabled OS to see the CPU as a pair, rather than a single CPU, but the effect is still the same, the single CPU is as around 90% as fast as a true pair of similar speed MP enabled CPU's.


Ban, I would assume you are talking about SUN and similar types of specialised CPU.

I used to run a SUN Ultra 20 Dual CPU machine, It was a server for my network about 2 years ago, it was outperformed by all other machines except the real cheap and crappy CPU's, like Cyrix. The machine was extrememly stable, Solaris is a great OS, I even ran Solaris for X86 on a number of Intel machines and saw a large performance boost over winblows, however the lack of driver support from other hardware vendors made this a short term experiment I'm afraid.

 

[ban-all-sheds]

Ban, I would assume you are talking about SUN and similar types of specialised CPU.

I used to run a SUN Ultra 20 Dual CPU machine, It was a server for my network about 2 years ago, it was outperformed by all other machines except the real cheap and crappy CPU's, like Cyrix. The machine was extrememly stable, Solaris is a great OS, I even ran Solaris for X86 on a number of Intel machines and saw a large performance boost over winblows, however the lack of driver support from other hardware vendors made this a short term experiment I'm afraid.

I'm not talking about Sun's SPARC chip (which BTW is now dead in the water given Sun's recent announcement that they have abandoned UltraSparc V), but another, and far superior, RISC processor.

Curious that you describe chips which run Unix, and Linux, as "specialised" - is that just because they don't run Windoze? (Solaris, BTW, is a "great OS" only in comparison to the latter - stacked up against other Unix OSs it is sadly lacking in scalability, reliability and functionality.)

And I'm not surprised that your Ultra-20's were easily outperformed - that's a very old system - US-II chips, I believe, clocked at 166-300MHz.

 

[AdamW]

Oooh, I think Ban must work for ARM then! Either that or Motorola/IBM/Apple with the Power PC units.

I am one who uses whatever OS he is given (I still use VMS at times) but choosing one for the home I go for the Ford Focus of operating systems. It's slow, pretty boring, not the best thing in the world to look at. But it runs anything I want it to, and anyone who comes round mine and says "mind if I check my e-mail?" can without a tutorial in Linux first.

Then there is the final bonus: there is no disagreement on how Windows should be pronounced! I favour the pronunciation of Linux used by it's creator, Linus Torvaldes, i.e. "Linn-Ucks". But soooooo many people insist on calling it "Line-Ucks". Nearly caused fist fights in my old student flat.

There's some fuel for more debate