Transmeta (TMTA)-The Monster That Could Slay Intel Message Board

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Transmeta (TMTA)-The Monster That Could Slay Intel

Public Reply Prvt Reply Mark as Last Read File

Previous 10 Next 10 Previous Next

To: Artslaw who wrote (312)	11/13/2000 10:59:43 AM
From: axp	Read Replies (2) of 421

i23jump said: McCarron points out that, in the past, companies such as Digital (now owned by Compaq) tried the software emulation approach with its Alpha processors but ran into the same performance degradation problems. Steve Walstra said: Digital made software emulators (called something like X86!) and binary compilers. The first is an application that can decode X86 binaries and lets you run the applications on an alpha, the second is a program which translates X86 binaries into native Alpha binaries (so you can run them directly). Neither one has anythinig to do with Transmeta, where all the "emulation" is all done on chip. The statement is just plain ridiculous--they are night and day different processes. The digital product was called FX!32 and was being integrated into Windows 2000 in cooperation with MS until Compaq pulled the plug on the whole Alpha NT platform. FX!32 enabled x86 programs to run on Alpha NT. The translation was only performed on the code that was actually executed and it did not produce a stand-alone alpha binary. You still needed to run the x86 program and FX!32 would switch to the translated portions where available. Transmeta has taken it a step further since they do it on the chip while FX!32 did it purely in software. But the concept is the same. In fact Transmeta has been recruiting and hiring ex-FX!32 developers precisely because of their experience directly applies. The primary point of the statement is that it's difficult to get good performance running x86 programs if you're doing it by translating to an underlying CISC instruction set. The reference to FX!32 is appropriate and IMHO the writer is correct. FX!32 only made sense while Alpha had a 2-to-1 performance advantage over x86 chips which enabled it to run x86 programs at approximately the speed of x86 chips. As the x86 MHz ramped up in 1996-1999 to exceed the speed of an Alpha it erased most of the advantages of an Alpha chip. I think Transmeta will be facing a similar "why should I bother" argument with their emulation/translation scheme.

Report TOU Violation

Share This Post

Public Reply Prvt Reply Mark as Last Read File

Previous 10 Next 10 Previous Next