lkml.org 
[lkml]   [2000]   [Jan]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectTM3120 and udelay.
Date
From

Hi Linus,

Do you mind me asking a few TransMeta questions? (Do you mind
answering them ;-)

How does Linux on the TM3120 (and TM5400) cope with the variable speed
in which the udelay loop is executed?

You are sometimes listed as the "mobile-linux" author and sometimes as
someone who has worked on the code-morpher. My guess is that you spent
the last few weeks on mobile-linux, but you actually worked on the
code-morpher most of the time. Right?

As minitiarization is all about reducing component count, would it be
possible for an OEM to fit a 2M flashchip, and have you move the
contents of the flash into write-protected RAM before starting the
x86 boot?

You guys managed to fit a complete x86 interpreter/compiler/optimizer
in 64k instructions (1mbyte / 128bit/instruction)? Aaargh! that's
lean! How much of that 64k instruction space is actually used? Does
the TM3120 have a 16bit PC, or is it 32 bits wide? (Or something
else?)

Registers are 32 bits wide right? That means that emulating Merced (or
whatever Intel came up with for a name this time) is left to a future
Transmeta CPU, which internally has the 64-bit support.

Intel CPUs allow me to upload new microcode (signed by Intel) using a
special instruction. Can I upgrade the Code-morpher on a transeta CPU
in a similar way? (Expected answer: Yes, but you're not allowed to
tell me).

As the TM3120 and TM5400 have serious hardware support for certain x86
features (e.g. the flags register and the MMU), it won't be easy to,
say, write a good & fast 68000 code morpher. Right?

But if say Motorola would show up wanting around 100M of your
processors a year, Transmeta would simply add the neccesary constructs
to the hardware (at say 2-4% silicon overhead) and run 68k code pretty
efficiently before the year is over... Right?

What is the voltage range for the CPU? The examples state that you get
cubic power-reduction. The examples name "clock at 90%", voltage at
90% -> .9^3 power consumption. However this becomes more impressive at
50% -> .5^3 power consumption: only about 10% of full-power. However,
the voltage range isn't that large, so my guess is that the CPU will
run at 90% core voltage at 90% of its rated frequency, and below that
you are only reducing the clock and get a linear power-advantage just
like everybody else. Right?

Roger.

--
** R.E.Wolff@BitWizard.nl ** http://www.BitWizard.nl/ ** +31-15-2137555 **
*-- BitWizard writes Linux device drivers for any device you may have! --*
"I didn't say it was your fault. I said I was going to blame it on you."

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:56    [W:0.058 / U:1.284 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site