You are not logged in. Please login or register.
Pages 1
On the same frequency 550 in CPUMark99 (32-bit):
K6-2 + 550 = 44.1
Duron 550 = 46.9
And after all k6-2 + on old boards Socket7 works.
Though it is no wonder. In 6-2 + as much a cache as in Durone, also it works on frequency of the processor (550).
Here so!
Xeon
And where you 550 found?
It was not necessary to search. The acquaintance dispersed old D-600, and for me tried to deliver it on 550 (100x5.5).
And you in casually did not ungear Djuronu?
You think Djuron FSB 200
And at 62 + FSB 100
KESH DID NOT UNGEAR!!!
And and here FSB. CPUMark measures only processor parrots, instead of system storage-processor. Though I will agree that influence it to render on results should. And frequency of the processor one and .
Adding from 08-10-2001 00:22:
P/s:
Loki13 , likely I got excited. Probably test also works on system the processor-memory, but here in what a piece. Speed of operation of storage both there, and there like the identical should be - 100 (if to consider that Duron works in the synchronous mode 200/100)
Well on CPU Mark and idle time 6-2 is equal Pentium III ... Anything surprising: and earlier at AMD good was...
Loki13
"Djuron FSB 200":eek:: lol:
Prediction gift?
At Durona 100 !!!!
Ve-For
100 reference... DDR... 200 resultant just 
Xeon
K6-2 + 550 = 44.1
Duron 550 = 46.9
And what the surprising? By operation on one frequency on practically at all processors approximate parity at an identical cache.
k6-2 112x5
cpumark-34
pumark measures the INTEGER part of processors.
ixbt.com measured Duron and 6-3 +, and 6-3 + on a ball floor Duron.
Korzh
I know! But Durona the bus 100, and about nobody spoke, and on 133 = FSB100...
Ve-For
At all 7 bus DDR generally. I not about storage...
And want the test where 6-2 + even tore Duron at equal frequency (truth 600 MHz, instead of 550)?...:eek: Please:
[img] http://www.ixbt.com/cpu/k6-2plus/calculations.jpg [/img]
And nevertheless it does not mean at all that 6-2 + more bright percents, rather than Duron...
P.S. Morals: it is necessary to be fond of synthetic tests less...
P.P.S.... And generally at a computer to sit...
Here thetas of anything surprising as integer at series 6 is shorter, than at 7. And on integer operations with the bad code optimization under a flow they, with other things being equal, will be faster. Yes, at systems on 6 productivity of a subsystem of storage is much worse, than at remaining modern platforms, without looking at frequency of the bus. Here, on it, even in the resulted test small lift-off 7 but if the task would get into a cache completely would be on the contrary is watched.
.... The Main lack of series 6 was not FPU, here, just, the considerable lag from the modern platforms on the optimized tasks (and today all such) is watched. But on majority FPU of operations series 6 fast, . does identical operations 6 for a smaller amount of clock periods, than 3 and 7.
..... It is a pity To me that these series 6 are laid off, I for technology equipment practically do not have changeover, well, unless , but it is necessary to alter and rewrite all.
Dmitry
So on idea it is possible to rake up second-hand. Now just the people on the sly
Korzh
About that at Athlon and Duron the system bus 100z I do not agree.
Look WCPUID.
There it is written:
System Clock - 100MHz
System Bus - 200MHz DDR
Even in descriptions of decryption of processors (the first line) it is written that the last letter A or B - 200MHz FSB, a C - 266MHz FSB.
P.S. By the way, how many time with you corresponded, but only now noted that you from Krasnoznameska. I too there earlier cores. Now in Moscow. , -2...
Xeon
About that at Athlon and Duron the system bus 100z I do not agree
Write 200 because it DDR!
Correctly:
100MHz DDR = 200 MHz.
DDR = Double Data Rate.
Important not number 100 or 200 it and at 200 on hundred kol-ebletsja. Important that at the data in 2 times it is transferred!
Therefore - 200, 266, the Stub 4-400.....
Adding from 10-10-2001 19:47:
And you still forgot mother ( for an offtopic ) But this TestTsPU on my Chejjntehe in just the same percents showed less than at it in a database - and in 4 variants different. And Vindovz during that moment can at you braked or that hung...
No! reviews with tests of these . At them and should be.
P/S. I understood that I in vain opened this subject. Let's it end.
IN WITH !!! ABOUT L AND AND !!!
Pages 1
Random topics