SPONSORED LINKS

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: TP 701 Upgrade by PEP



At 09:11 AM 11/3/97 -0800, Mitchell Yee wrote:
>     A couple of offers:
>     
>     1.  I use a statistical analysis and reporting package called SAS.

	Actually, that's *the* stats package--I know of it, but I've never used it.
	
>     Randall, if you're interested and willing, I'll be happy to ship you
>     a hard drive with W95B and SAS loaded.  You can then run a few
>     analysis programs on a small database, then a large database, then
>     ship the disk back to me - I'll cover the shipping costs.  Don't worry
>     about the disk - it's a spare...

	If you include the benchmarks you want me to run *on* the disk too, that
would be useful.

	You'll also need to send along a page of "instructions" concerning what
you want me to do as far as running the datasets through SAS.  I am
completely unfamiliar with its use.

	And I presume you'll just run it on your machine first, then send it to me
after I get the upgrade done on my 701?

	Mitchell, would you be kind enough to pull together the *definitive* set
of benchmarks that I should be running?  There have been a whole host of
suggestions, and right now the only thing I would definitely say I'll plan
on doing is Winbench (stripped down version at 4 MB) & Wintune--perhaps
Winstone, but I don't like the idea of going through that big CD-ROM with
its possible overhead (I would also have to *get* the CD-ROM from someone,
as well as the one for Winbench if the full version is desired).

	I just want to know that everyone can agree on which test(s) will provide
them the picture they need to make a judgement about the upgrade's
performance.

	I also don't want to have a veritable *crapload* of different benchmarks
clogging up the works--lets keep it to 2 or 3, please (plus whatever ones
the OS/2 folks want--and perhaps your SAS statistics test--that's actually
a good indicator since statistical functions can be a good indicator of
processor speed).

	I also want to avoid wasting time on anything that is sub-system
specific--lets just see (1) how fast CPU-intensive chores are compared to
the "norm" and (2) how much more speed you can expect from everyday chores
(a big document in Word, a big spreadsheet in Excel, etc.).  There's no
sense in trying to see how fast the Hard Drive is, etc.--Just the CPU and
how the whole system performs with the new CPU.

>     2.  I'll be happy to collect and tabulate and produce a Web/text page
>     with any benchmarks reports that are sent to me, perhaps in 
>     collaboration with ???

	In collaboration with Emanuel Brown would be great.  I'll E-mail all
results to both of you for putting together, etc.

>     Here's a suggested list of data points to collect when reporting
>     benchmarks.  Since this is just a start, suggestions and changes are 
>     welcome:
>     
>     1.   Machine model and type:        2640-4UG
>     2.   CPU:                           486DX75
>     3.   Hard drive size:               2.1 GB

	Just have the stock 720 HD in the 701.

>     4.   RAM size:                      24 MB
>     5.   BIOS version:                  something like BTU03
>                                        
>     6.   OS (version):                  Win95B (OSR2)

	It'll have to be "normal" Win 95.

>     7.   Disk system:                   FAT32

	It'll have to be FAT16.

>     8.   Video settings:                256 colors
>     9.   Virtual disk settings:         64 MB (fixed)

	May not have this much space to work with, but I'll see what I can do.

>     10.  Special machine configuration: -none-
>     
>     
>     11.  Benchmark test:                Wintune
>     11a. Report:                        xxxxxx
>     
>     12.  Benchmark test:                Winbench
>     12a. Report:                        xxxxxx

	I get the idea...


-------
Randal J. Whittle		whittle@usc.edu	(213) 740-7775
Director, Electronic Commerce Program
Marshall School of Business, University of Southern California