Re: Atmel releasing FLASH AVR32 ?


"Ulf Samuelsson" <ulf@xxxxxxxxxxxxx> wrote in message news:ev2i1h$qhh$1@xxxxxxxxxxx

Multithreading on a high end general purpose CPU gives problem on their own.
Especially with cache trashing.

Absolutely. The "solution" is to add more cache...

With an embedded core where you use tightly coupled high bandwidth memory
for most of the threads you do not have that problem

Same solution: more fast on-chip memory.

I think it is eminently useful for assymetric multiprocessing where
you have some dedicated tasks to do which are best implemented
in a separate CPU to avoid real time response conflicts and can
be implemented in a low end 32 bitter.

I'm not quite sure what you're saying here. Are you advocating
asymmetric multiprocessing or asymmetric multithreading?

I think you need to stop trying to explain why a single CPU
is better than a multiththreaded CPU, because noone is
using a single CPU for implementing two simulaneously
operating software MACs.

First of all, you're the one that claims one CPU is better than 2...
I believe 2 CPUs is better in many cases - multicore is the future.
However if you do move to a single (faster) CPU then it doesn't
make much difference in terms of realtime response whether that
CPU is multithreaded or not. You seem to believe that threads are
somehow much better than interrupts - but as I've shown they are
equivalent concepts.

If you continue, that just proves that you are either ignorant or not listening

That kind of response is not helping your case. If you believe I'm wrong,
then why not prove me wrong with some hard facts and data?

The issues is replacing multiple CPUs/Memory Subsystems
with a single multithreaded CPU addressing a memory subsystem´
consisting of internal TCM memory, internal loosely coupled
memory (flash?) and external memory.

Most realtime CPUs have some form of fast internal memory,
this is not relevant to multithreading.

Eight cores and 16 threads (probably they mean per-core?) is impressive
for what sound like fairly mainstream cores.

It clearly says 2 threads per core. Any more would be a waste.


Look at Sun and UltraSparc T1, they certainly do not see the boundaries that you see.

The T1 has tiny caches and stalls on a cachemiss unlike any other
high-end out-of-order CPU, so they require more threads to keep going
if one thread stalls. It is also designed for highly multithreaded workloads,
so having more thread contexts means fewer context switches in software,
which can be a big win on workloads running on UNIX/Windows (realtime
OSes are far better at these things).

I do not think that they are limited by Intels vision...
Also I pointed you at the new MIPS Multithreading core.
They certainly do not agree with You!

If you do not understand the differences between cores like Itanium-2,
Pentium-4, Nehalem, Power5, Power6 (all 2-way multithreaded),
and cores like the T1, MIPS34K and Ubicom (8+ -way threaded),
then you're not the expert on multithreading you claim to be.

Wilco


.

Relevant Pages

  • Re: a dozen cpus on a chip
    ... At this moment, my PC is running 389 threads, on one CPU. ... as the different cores would use their own memory and now you have to ... move data all the time between cores.. ... The rest can and should be done in one core task switching. ...
    (sci.electronics.design)
  • Re: OT Dual core CPUs versus faster single core CPUs?
    ... The real use for multiple cores will to be to assign one function per ... specific device drivers, file managers, TCP/IP socket managers, things ... How do you think a given CPU ... The OS cpu will assign it a task, create its memory image, set up its ...
    (sci.electronics.design)
  • Re: server upgrade question
    ... GIG of memory running Debian 4.0. ... servers into 5 servers. ... The new server will have 16 cores with 64 GIG ... Imagine you have two computers with one CPU, ...
    (Debian-User)
  • Re: multibyte subtraction using AAS and ASCII Decimals
    ... earlier 8-bit CPU updated to 32-bits. ... This reduced transistor counts. ... count so that you could pack all the cores onto a modern die. ... this could allow more memory sharing). ...
    (alt.lang.asm)
  • Re: Threading NOT working as expected
    ... doesn't mean a single thread is doing work at any time. ... Most of the time in a modern system is spent waiting for I/O and memory ... There are some very special cases where the CPU is actually ... Multithreading won't help for memory fetches. ...
    (comp.lang.perl.misc)