[rc5] Re: Supercomputers & Bovine, Q&A

Ryan Dumperth woodie at indy.net
Mon Aug 4 21:50:58 EDT 1997


At 07:53 PM 8/4/97 -0500, Eric P. Anderson wrote:
>On Mon, 4 Aug 1997, Murray Stokely wrote:
> 
>> It probably would have been easier to point out that his entire
>> argument that RISC is better was pointless especially in this
>> context.  CISC machines such as x86's are outperforming RISC machines
>> such as alphas that run at over twice the clockspeed.
>
>This is nonsense.  Please show me an x86 chip that can outrun a 600
>mhz Alpha. 

Taken from <http://www.llamas.net/rc5/speed.html>

NT     Digital DPWS 433a  Alpha 21164 at 433 Mhz  485000 keys/sec
NT 4.0 Intel              Pentium Pro at 233 Mhz  546855 keys/sec

--------

At 07:38 PM 8/4/97 -0500, root <marcus at dfwmm.net> wrote:
>bennett_t1 at popmail.firn.edu wrote:
>> 
>> True, RISC chips are faster than CISC, and therefore better, In most
>> respects.
...
>
>However, RISC means more instructions fetched to do a certain task.
>That means bigger programs that occupy more memory, more disk space,
>more cache and ultimately more power consumption because the poor
>chip has to drive all of those external address and data lines more
>often.
>

Take from
<http://www.mot.com/SPS/PowerPC/library/press_releases/750announcepr.html>

PowerPC 604e at 350MHz:
•Processor Size: 47 sq. mm
•Power Consumption: less than 8.0 watts @ 350 MHz
•Manufacturing Technology: .25-micron
•Performance at 350 MHz (est.): 14.6 SPECint95 and 9.0 SPECfp95

Taken from <http://www.intel.com/pressroom/archive/releases/dp5797fs.htm>
•Processor Size: 202 sq. mm
•Power Consumption: less than 43.0 watts @ 300 MHz
•Manufacturing Technology: .35-micron

Taken from: <http://www.intel.com/procs/perf/PentiumII/spec95int.htm> and
<http://www.intel.com/procs/perf/PentiumII/spec95fp.htm>

•Performance at 300 MHz (est.): 11.6 SPECint95 and 7.2 SPECfp95

Performance versus power consumption seems to favor Motorola/IBM's RISC
implementation over Intel's RISC/CISC masterpiece quite handily.


----
To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the body.



More information about the rc5 mailing list