Re: Chinese Alpha?

On May 3, 12:38 pm, ChrisQ <m...@xxxxxxxxxxx> wrote:
On 05/02/12 01:12, Michael Kraemer wrote:



I don't know what feeds your pipe dreams,
but the Alpha never had a good track record in power efficiency.
In fact, its inability to adapt to the needs of embedded devices
(with the exception of AXPvme, which failed miserably)
was one reason for its demise.

Sounds like rewriting history ?. I ran alpha machines for a decade
or more and don't remember them being any more power hungry than
equivalent x86 machines of the time, though they were much, much
faster.

With shrinks and improved process technology, i'm sure that Alpha
would have been more than competitvve with current designs, just
as ibm power is now for it's intended market.

Wonder how fast the last of the microvaxen would have been with
similar upgrades ?...

Regards,

Chris

"I ran alpha machines for a decade or more and don't remember them
being any more power hungry than equivalent x86 machines of the time,
though they were much, much faster."

Alphas weren't notably more power hungry than comparable x86 from the
same era. I'll offer an EV4 and an EV5 system as examples.

The EV4 example is the AlphaStation 400, which was a EV4/21064 (at up
to 233MHz if I remember rightly). The very same box and the very same
power supply and cooling was available with different part numbers and
processor daughterboard as either a high(ish) end x86 desktop (the
Celebris XL, which went all the way up to a dual Pentium 166 or a
Pentium Pro) or an entry level AlphaStation (or with the same
hardware but different software and licensing, as an AlphaServer).

The EV5 system I'm thinking of was the Personal Workstation (Miata
etc) with an EV5 at up to 600MHz, again sold in NT-only and more
expensive Unix/VMS-licenced versions. These again shared a box and
power supply and pretty much everything except the processor
daughtercard (in industry standard but not widely used NLX format)
with their x86 equivalent.

I have an unconfirmed suspicion that something similar went on with
the AlphaServer 1000 (Mikasa) but I can't remember the name of the
corresponding PCBU/Intel server box. Supporting evidence (or
correction) welcome.

These were not cheap bog-standard low-end systems, either in x86 or
Alpha flavours, but the NT versions of the Alpha systems weren't
wildly different in price from their x86 relatives, and performance-
senstive customers I dealt with were happy to pay a little extra for
the extra performance, and those using them for VMS or Tru64 were also
happy that they had an affordable reasonable-performance VMS/UNIX box,
quite competitive with other RISC boxes of the era (though courtesy of
generally relying on "the channel" to do marketing and promotion, not
many prospective customers actually knew this).

So, on the subject of power consumption, I'd say Chris's recollection
is perfectly reasonable and supported by the evidence.
.

Relevant Pages

  • Re: Alpha, beta, power, CL, CI
    ... Alpha is the probability that our test would conclude that T is better than S even though this is not the case. ... Beta and power are two ways of talking about the same thing. ... They are all about how likely it is that, given fixed alpha, the number of patients N we will test the treatments on and assuming a scenario under which T actually is better than S, our test will actually conclude that T is better than S. That probability is called power, while beta is simply 1-power. ... Confidence intervals corresponding to a certain confidence level are loosely speaking* based on the following idea: Given outcomes to our experiment, we want to quantify how precise our estimates of e.g. for the difference in response rates between the two treatments are, for this purpose we want to specify a range about which we are confident that the true difference will lie in it. ...
    (sci.stat.math)
  • Re: VMS backup: competition from Apple
    ... What will a desktop PC do with 32 cores? ... absolutely nothing to do with competing with Alpha. ... Sun did multi-core because it saved power, ...
    (comp.os.vms)
  • Re: Chinese Alpha?
    ... lead the market with power efficient Alpha, ... but the Alpha never had a good track record in power efficiency. ... When talking multiple cores, the "glue" on the chip, memory controllers, interprocessor communications, and such in EV7 and EV7z, even with the larger die size was doing things that out performed competing CPUs that had continued development and smaller die size. ... then the majority just might be satisfied with VMS running on x86-64. ...
    (comp.os.vms)
  • Re: Who "invented" the null hypthesis?
    ... The probability of a Type I errror is called alpha ... RU> Clinical research designs typically compute the type 2 power ... RF> This is complete gibberish in responding to someone who asks ... (Bob isn't really ...
    (sci.stat.math)
  • Re: controversial paper
    ... > aren't those machines any better than Pentium machines? ... SpecFP results and those of the fastest x86. ... As someone who's programmed at a low level for x86, Sparc, HPPA, Alpha, ... Power, TI-C80, I can genuinely say that I find Alpha, Power, ...
    (sci.crypt)