[Brianflys]

On a previous forum, Spenser wrote:

"So far as I understood from a German PDA-manufacturer, Intel has inofficially (= interdepartmentally) confirmed a microcode bug concerning memory move which is causing so much hassle and ranting. He added, that some products may even have to be returned to the manufacturer for "repair" (=replacement of the xscale-proc) and that at this moment it isn't even clear if the others can be "healed" with a SW-patch. He was sure though that Intel will deliver updated xscale processors later this year. So I can only advise you not to buy an xscale-product now but to wait till Oct-Nov. And yes, there is also the rumour that Microsoft will deliver a patch in the same timeframe for a better xscale-support of PPC 2002."

My question is...has this been corrected with the new Xscale processor in the iPAQ 5450? Or are we stuck with the same issue?

[Brianflys]

On a previous forum, Spenser wrote:

"So far as I understood from a German PDA-manufacturer, Intel has inofficially (= interdepartmentally) confirmed a microcode bug concerning memory move which is causing so much hassle and ranting. He added, that some products may even have to be returned to the manufacturer for "repair" (=replacement of the xscale-proc) and that at this moment it isn't even clear if the others can be "healed" with a SW-patch. He was sure though that Intel will deliver updated xscale processors later this year. So I can only advise you not to buy an xscale-product now but to wait till Oct-Nov. And yes, there is also the rumour that Microsoft will deliver a patch in the same timeframe for a better xscale-support of PPC 2002."

My question is...has this been corrected with the new Xscale processor in the iPAQ 5450? Or are we stuck with the same issue?

[rave]

Could this be the article?:

"Der Benchmark-Test ist nun upgedated. Signifikant hat sich nichts geändert. Eines jedoch scheint sich immer mehr herauszukristallisieren: Die Memory-Move-Funktion des Prozessors, die den Gerüchten nach beim XScale durch einen Bug beeinträchtigt wird, ist in den Tests sowohl beim LOOX als auch beim 39xx signifikant langsamer als bei den StrongARM-Geräten..."

(translation) "The bench mark test is now upgedated. Significantly nothing changed. One however seems to out-crystallize ever more: The MEMORY move function of the processor, which is impaired the rumors after with the XScale by a nose, is significantly slower in the tests both with the LOOX and with the 39xx as with the StrongARM devices..."

I asked because I found nothing severe here. It may be true that memory moves on XScales are a bit slower than SA's, but it's by design, and far from being a CPU microcode bug.

Or perhaps there is another source?