Kenny1702
2003-10-31, 23:27:54
Ich hab mich an nem alten Artikel (http://www.3dcenter.org/artikel/2002/10-02_a.php) versucht, es gibt z.Z. noch bestimmt massig Kommafehler und der Satzbau ist manchmal etwas "komisch" bzw. die Sätze sehr lang, aber es sollte erstmal sehr nahe an Leos Text liegen. Seite 2 werde ich vielleicht morgen nachreichen. Ich hoffe niemand mockiert sich über mein nicht abgesprochenes Vorgehen;).
Nun die Übersetzung soweit:
Athlon XP with 166MHz FSB
October 2nd 2002 / by Leonidas, translation by Kenny1702:D / page 1 of 2
On yesterday's October 1st 2002 AMD presented the first processors in shape of the Athlon XP models 2.16 GHz / 2700+ and 2.25 GHz / 2800+ on base of the otherwise unchanged Thoroughbred-B core.
By that one gives the long expected leap of a higher front side bus a try, as some time has gone by since the last leap from 100 to 133 MHz DDR FSB (Athlon Thunderbird 1.33 GHz in March 2001).
As expected the new front side bus shows itself only with a few percent in performance.
Through the concurrent clock leap of the 2.25 GHz / 2800+ model it is enough to reach again a tight first place in favor of AMD and in disadvantage of the till now leading Pentium 4 Northwood with 2.8 GHz, even if this one runs with PC1066 RAMBUS-Memory.
Recomandable reviews from a subjective point of view are the ones of Computerbase (very extensive), of Hard|OCP (comparison KT400- to nForce2-base) and of Ace's Hardware (good benchmark selection), a complete list of all reviews can be found at the end of this article.
But the good results of the new Athlon XP models were mainly reached on nForce2-bases, which are supposed to appear in stores not before this month.
AMD had provided with foresight all testers of the new processors with nForce2-mainboards, which explains the great amount of tests that are based on nForce2-mainboards.
The nForce2 reaches indeed mostly slightly higher results as VIA's KT400, like the test of Hard|OCP shows for instance.
But also with a KT400-base AMD's model rating compared with the Pentium 4 can be still kept with the new AMD processors.
Why the nVidia nForce2 presents better results than the VIA KT400, will have to be explained.
Perhaps the oversized bandwidth of the nForce2 through its dualchannel DDR333 memory interface (5,0 GB/sec) has some effects, although the Athlon XP itself on FSB166 has only a maximum bandwidth demand of 2,5 GB/sec.
Perhaps nVidia has just only delivered a faster chipset, while it first requires an optimisation of the KT400 towards the KT400A for VIA.
This possibillity is indicated through the slim difference between the singelchannel and the dualchannel mode of the nForce2, which x-bit could find out.
Likewise has the phenomenon, that some KT333 mainboards profit only a little bit or not at all from the front side bus of 166 MHz DDR, while others show performance improvement of round about 5 percent at otherwise same processor clock, also to be examinated through future tests.
Concerning this Hard Tecs 4U published an interessting article that go into this subject, which is hopefully fixable through an appropiate BIOS update of the mainboard manufacturer.
Interessting are the reached overclocking results of the new procesors for sure, since the 2.13 GHz / 2600+ model already reached good 2.3 up to 2.5 GHz when overclocked.
The articels of the Athlon XP 2700+ and 2800+ so far could find out following:
ascertained stable overclocking results
(at CPU voltage)
average
(clock gain)
The Thoroughbred-B core ist not yet finished, but it doesn't seem to exist much clock elbowroom any longer.
The reached overclocking results of on average 170 MHz more clock are compared to the processor clock (less than 10 percent overclocking) more unrelevant and show no real overclocking suitability for the Athlon XP 2700+ and 2800+ models.
One, who wants to realise real overclocking successes of 20 or more percent will furthermore only have luck with the low clocked Pentium 4 models.
Apart from the performance affair the new front side bus probably also raises new compatibility problems.
In detail the new processors with 166 MHz DDR FSB (brief FSB166) now require mainboards that have on one hand the clock possibility regarding the FSB and offer on the other hand the required PCI and AGP divisors of 1/5 and 2/5.
Until now have nVidia through the nForce2 chipset, SiS through the 746 chipset and also VIA through the chipsets KT333 and KT400 confessed to the official support of FSB166.
nVidia had done this, as everybody knows, already before the launch of the FSB166 processors and by that given a hint to exactly this launch, at VIA one had kept back concerning this and released KT333 and KT400 until two days ago officially only for 100 and 133 MHz DDR FSB.
Since yesterday the KT333 as well as the KT400 product pages describe now these chipsets FSB166 suitable, although the corresponding block diagram would still require concerning this a revision.
All other chipsets are officially not FSB166-able.
However it also depends with older chipsets on the particular mainboard, if it can made going with FSB166 or not.:
officially FSB166 suitable ----- 166 MHz DDR FSB adjustable ---- PCI divisor 1/5 and AGP divisor 2/5 available
no ---- on good retail mainboards ---- on good retail mainboards ( most nForce mainboards offer the possibility to fix the PCI and AGP clock on 33 resp 66 MHz.
yes ---- yes ---- yes
no ---- on good retail mainboards ---- on good retail mainboards
no ---- on good retail mainboards ---- on good retail mainboards
no ---- on good retail mainboards ---- possibly
yes ---- on good retail mainboards ---- yes
no ---- on good retail mainboards ---- no
no ---- on good retail mainboards ---- no
yes ---- on good retail mainboards ---- officially yes ( it is not sure, if every mainboard facturer has disclosed the divisors)
yes ---- yes ---- yes
Nun die Übersetzung soweit:
Athlon XP with 166MHz FSB
October 2nd 2002 / by Leonidas, translation by Kenny1702:D / page 1 of 2
On yesterday's October 1st 2002 AMD presented the first processors in shape of the Athlon XP models 2.16 GHz / 2700+ and 2.25 GHz / 2800+ on base of the otherwise unchanged Thoroughbred-B core.
By that one gives the long expected leap of a higher front side bus a try, as some time has gone by since the last leap from 100 to 133 MHz DDR FSB (Athlon Thunderbird 1.33 GHz in March 2001).
As expected the new front side bus shows itself only with a few percent in performance.
Through the concurrent clock leap of the 2.25 GHz / 2800+ model it is enough to reach again a tight first place in favor of AMD and in disadvantage of the till now leading Pentium 4 Northwood with 2.8 GHz, even if this one runs with PC1066 RAMBUS-Memory.
Recomandable reviews from a subjective point of view are the ones of Computerbase (very extensive), of Hard|OCP (comparison KT400- to nForce2-base) and of Ace's Hardware (good benchmark selection), a complete list of all reviews can be found at the end of this article.
But the good results of the new Athlon XP models were mainly reached on nForce2-bases, which are supposed to appear in stores not before this month.
AMD had provided with foresight all testers of the new processors with nForce2-mainboards, which explains the great amount of tests that are based on nForce2-mainboards.
The nForce2 reaches indeed mostly slightly higher results as VIA's KT400, like the test of Hard|OCP shows for instance.
But also with a KT400-base AMD's model rating compared with the Pentium 4 can be still kept with the new AMD processors.
Why the nVidia nForce2 presents better results than the VIA KT400, will have to be explained.
Perhaps the oversized bandwidth of the nForce2 through its dualchannel DDR333 memory interface (5,0 GB/sec) has some effects, although the Athlon XP itself on FSB166 has only a maximum bandwidth demand of 2,5 GB/sec.
Perhaps nVidia has just only delivered a faster chipset, while it first requires an optimisation of the KT400 towards the KT400A for VIA.
This possibillity is indicated through the slim difference between the singelchannel and the dualchannel mode of the nForce2, which x-bit could find out.
Likewise has the phenomenon, that some KT333 mainboards profit only a little bit or not at all from the front side bus of 166 MHz DDR, while others show performance improvement of round about 5 percent at otherwise same processor clock, also to be examinated through future tests.
Concerning this Hard Tecs 4U published an interessting article that go into this subject, which is hopefully fixable through an appropiate BIOS update of the mainboard manufacturer.
Interessting are the reached overclocking results of the new procesors for sure, since the 2.13 GHz / 2600+ model already reached good 2.3 up to 2.5 GHz when overclocked.
The articels of the Athlon XP 2700+ and 2800+ so far could find out following:
ascertained stable overclocking results
(at CPU voltage)
average
(clock gain)
The Thoroughbred-B core ist not yet finished, but it doesn't seem to exist much clock elbowroom any longer.
The reached overclocking results of on average 170 MHz more clock are compared to the processor clock (less than 10 percent overclocking) more unrelevant and show no real overclocking suitability for the Athlon XP 2700+ and 2800+ models.
One, who wants to realise real overclocking successes of 20 or more percent will furthermore only have luck with the low clocked Pentium 4 models.
Apart from the performance affair the new front side bus probably also raises new compatibility problems.
In detail the new processors with 166 MHz DDR FSB (brief FSB166) now require mainboards that have on one hand the clock possibility regarding the FSB and offer on the other hand the required PCI and AGP divisors of 1/5 and 2/5.
Until now have nVidia through the nForce2 chipset, SiS through the 746 chipset and also VIA through the chipsets KT333 and KT400 confessed to the official support of FSB166.
nVidia had done this, as everybody knows, already before the launch of the FSB166 processors and by that given a hint to exactly this launch, at VIA one had kept back concerning this and released KT333 and KT400 until two days ago officially only for 100 and 133 MHz DDR FSB.
Since yesterday the KT333 as well as the KT400 product pages describe now these chipsets FSB166 suitable, although the corresponding block diagram would still require concerning this a revision.
All other chipsets are officially not FSB166-able.
However it also depends with older chipsets on the particular mainboard, if it can made going with FSB166 or not.:
officially FSB166 suitable ----- 166 MHz DDR FSB adjustable ---- PCI divisor 1/5 and AGP divisor 2/5 available
no ---- on good retail mainboards ---- on good retail mainboards ( most nForce mainboards offer the possibility to fix the PCI and AGP clock on 33 resp 66 MHz.
yes ---- yes ---- yes
no ---- on good retail mainboards ---- on good retail mainboards
no ---- on good retail mainboards ---- on good retail mainboards
no ---- on good retail mainboards ---- possibly
yes ---- on good retail mainboards ---- yes
no ---- on good retail mainboards ---- no
no ---- on good retail mainboards ---- no
yes ---- on good retail mainboards ---- officially yes ( it is not sure, if every mainboard facturer has disclosed the divisors)
yes ---- yes ---- yes