Conference 7.286::macintosh

    I own a PM 7200/90Mhz, now with 32Mb RAM and no L2 cache and had a
    Centris 650 (68040/25Mhz) with 32Mb as well that I still maintain for
    my parents needs.
    
    The PM seemed equivalent to the Centris when running 68K applications
    (office applications and games), when they were both running MacOS v7.5
    and had about 16Mb RAM each.
    
    Then came MacOS v7.5.3 and revisions which I cannot qualify as I was
    more often rebooting both systems after they freezed than actually working.
    
    Now, using MacOS V7.5.5, and after upgrading to 32Mb, the PowerMac seems
    really faster (at last!) than the Centris.  And my mood is much better
    now ;-)  Although, to my opinion, the PM still would be barely 1.5 to twice
    as fast when running 68K software, and 2 to 2.5 times faster when running
    native applications, than the Centris.  No big difference for the extra
    bucks and all the roaming about Power power power PC...
    Magazines and advertisements mentionned factors as high as 4 or 5 times
    faster...  No way.  Either they lie or they may measure CPU raw power, not
    the whole system in usual work conditions.  Of course the configurations
    they use for benchmarks are enhanced a lot w.r.t. the configurations
    out of catalogs (the Centris came with only 4Mb, soldered on motherboard.
    The PM came with 8Mb and no L2 cache...  Almost useless systems as such
    I would say).
    
    
     
    I was told many times that SpeedDoubler would 

re:<<< Note 875.32 by AZUR::HUREZ "Connectivity & Computing Services @VBE. DTN 828-5159" >>>
                              -< No big deal... >-

>     Although, to my opinion, the PM still would be barely 1.5 to twice
>     as fast when running 68K software, and 2 to 2.5 times faster when running
>     native applications, than the Centris.  No big difference for the extra
>     bucks and all the roaming about Power power power PC...
>     Magazines and advertisements mentionned factors as high as 4 or 5 times
>     faster...  No way.  Either they lie or they may measure CPU raw power, not
>     the whole system in usual work conditions. 


    Your PowerMac 7200 is crippled by not having a Level 2 cache.
    It makes a big difference on compute bound tasks such as
    68K emulation.

    /Bob Anderson

    >> Your PowerMac 7200 is crippled by not having a Level 2 cache.
    >> It makes a big difference on compute bound tasks such as
    >> 68K emulation.
    
    Yeah...  I won't ever forgive Apple for having sold that machine
    with only 8Mb RAM and without L2 cache (OK, by those times, RAM
    was expensive, but the 7200 was really expensive as well...).
    
    I ordered a 256K L2 cache more than a month ago, and I'm still
    waiting for it...  Were it a Wintel PC, I would have got it right
    away at the local dealer, next door to mine (Ok, I may also have spent
    days in opening the case, disconnecting things to access the slot,
    and finally returning the incompatible module for exchange with a
    different one, still incompatible ;-)
    
    	-- Ol.
    
    
    

    Can anyone point me to some benchmarks or other comparisons?

    PowerPC 604 versys 601/603?

    Also, how would a 604 compare with Pentium or PentiumPro...
    what MHz Pentium would be about the same performance as a 604?

    When I look at Mac prices right now, they appear to be a lot
    lower than PC prices.  But maybe i'm not comparing same performance
    systems....
    bob

If you're considering a Mac, check out the nice MacUser pages on this at:

	http://www.zdnet.com/macuser/mu_0397/feature/whichmac/

For speed comparisons amongst Macs, click on the Rated CPUs option in
the narrow frame at the left, as well as the CPU Report Cards under
Related Sites.

Keep in mind that performance depends on much more than raw CPU speed.
Things like internal bus speed and the presence of a 2nd-level cache
can have near-first-order effects on performance.  (Still, as a rough
rule of thumb, the Power Computing page rates a 120MHz 604 as about
the same as a 180MHz 603e).  The sites I mention above use a standard
Macintosh benchmark suite for comparison, including measures for things
other than just raw speed.

    I thought the 604e is supposed to be roughly equal to the same Mhz
    PentiumPro. A 250Mhz 604e is going to be faster than any
    shipping Pentium.
    
       erik

    Current rule of thumb is 603e is on par with Pentium, 604e is on par
    with Pentium Pro at equal MHz.  But as was already mentioned, the other
    system components have a lot of effect on performance.  The memory
    bandwidth, L2 cache size, speed of graphics controller, disk seek and
    transfer speeds, 2x, 4x, 6x, 8x, 12x CD-ROM drive if you use many
    CD-ROMs etc...
    
    Here are some web pointers to benchmark locations.
    
    	<http://www.lava.net/~robart/Speedo.html>
    	<http://ng.netgat.net/%7Eengstrom/cc.html>
    	<http://www.netlib.org/performance.html/PDStop.html>
    	<http://www.zdnet.com/macuser/cpu/>
    
    I hope I typed them in OK.  I have not looked at these links recently
    so if some are stale, I'll say I'm sorry now.
    
    					Bob Harris

    
    Picked this up at: http://www.rimatech.com/html/ppcnews.html
    
    603ev/166 Trounces Pentium/133 (Nov 2, 1995) 
    
    603ev/166 trounces Pentium/133 on SPECint95 and Byte-int, even when
    comparing a slighly enhanced note book vs a
    desktop. 
    
    
    Proc.      Pentium   Pentium   604       604        603ev
    MHz        133       133       133       133        166
    System     Hot box   Micron    Hot Box   Carolina   Woodfield+
    L2         1M        256K      1M        512K       512K
    Memory     EDO       EDO       SDRAM     DRAM       DRAM
    
    SPECint95  3.68      -         4.9       4.55       3.93
    
    Byte-int             1.92                2.92       3.37 
    Byte-fp              1.63                2.97       2.40
    
    All based on preliminary measurements made by Dave Jaffe of PPS
    performance. Woodfield is a notebook, the system
    used here has added cache and more memory. Even better (+10-15%)
    results would be expected if the 603ev was placed in
    a performance optimized system. 
    
    If you want to see a comparison of lots of systems (not chips, systems)
    see the CINT92 SPEC Benchmark Summary at;
    http://performance.netlib.org/performance/html/new.spec.cint92.col0.html
    (BTW Our AlphaServer came in at the top and we have 17 out of the top
    25 slots (5 out of the top 10))
    
    For a comparison of several CPU's see;
    http://davinci.mechanik.th-darmstadt.de/ag3/bauer/cpus.html
    It's in table format, so I won't put the whole thing in, but;
    
    Proc	Ship	Clock	SPEC92	SPEC95
    essor	Date	Mhz	int fp	int fp
    P6		1996	166	293 261	7.3 6.2
    P6		1995	200	320 283	8.2 6.8
    P6		1996	200		8.7 6.7
    604e	1996	166		6.0 5.0
    604e	1996	200		8.0 6.3
    
    From these numbers it looks like the Pentium Pro (aka P6) is a bit
    faster than the 604e at the same speed on SPEC95. 

    Also see 902.23.
    
    				-John

    thanks!  this is pretty much what I've been looking for.

    At our school, we are looking to put in a network.  A guy
    in town in responsible for the networks at the company where
    he works, and has volunteered to lead a 'volunteer' effort
    to pull the wires to wire the school.  But he is pretty
    adamant about making the server on this network be Pentium NT.
    To him, the whole school should be PCs not these silly
    Macs and the server is Step 1.  He also arguing that we shouldn't
    bother putting ethernet cards into the older Macs in the
    school...throwing good money away.  And that any new machines the
    school buys should be PCs....

    the school is VERY Mac oriented....most of the teachers own a Mac,
    and none own a PC, some have their personal machine in their room.
    The school owns about 15 or so Macs...LCs and various Powerbooks.
    the kids and teachers are very used to them...the school has no
    PC software.

    Anyway..you get the idea....the guy is not getting a real warm reception
    in his crusade to convert to PCs.  But we really could put in an
    NT server and leave the clients all Macs.  Many of the arguments that
    workfor the 'clients' dont work for the server.  But part of his
    argument is that PCs are so much cheaper...that he can build a server
    for SO MUCH less than using a Mac.  I dont know enough to really know...
    when i look in various catelogs, I see LOTS of IDE CD Towers and tape
    systems for backup.  I see few SCSI ones, and they do cost more.
    the big SCSI disks do cost more.  But SCSI is not THAT much more than
    IDE.  I've heard talk that IDE is very unreliable (compared to SCSI)?
    You can buy all the software you need to build a server using NT
    for very cheap (Novell for $200, most of the rest free?).  And Apple
    charges a lot?

    Any thoughts on this....perhaps this is not the right topic.
    But I've been comparing PCs in PC catelogs and comparable
    MAcs....and I just dont buy the argument that PCs are cheaper....
    bob

This seems to be a typical PC argument that holds no water! In the 
case of our elementary school which is very Mac based we chose a Macintosh 
Server as it provides the ability to use the old machines wired via Appletalk
phone net and ethernet for both PC's and New Mac's.  The older machines can 
share Network resources such as printing,a network modem,file sharing and 
limited browsing. The newer machines will be hooked up to ethernet for the 
same uses. Don't forget electronic mail? The building has been using a phone 
net network for four years now and will be wired for ethernet on Net Day 97.
Each room has at least one high level multimedia machine an a couple of Mac
plus's,or SE's or LC's.

The biggest expense with PC's is support and training and in some cases certain
teachers will not make the transition . Quite frankly there are only three 
teachers out of 24 or so who use PC's. It should be noted that teachers were 
given the choice of platform and have clearly seen the Macintosh environment 
as much easier to get work done.

But in reality by giving the teachers the final choice is what made the 
decision at this school. It doesn't matter what this fellow has to say as it 
will be the teachers who have to support the network and machines after this 
guy is long gone. I was able to convince the principal of this who is really a 
PC user. 

Educational discounts from Apple use to pretty much equalize the PC to Mac 
dollar difference. 

As far as an Internet server the school bought a top of the line Power Mac,
which gives you several software options. The most expnsive is in the 
neighborhood of $400 and has everything from DNS,WEB,Mail server software to
HTML editors,data base,backup,dialup software etc. I can't imagine the package
not having anything you would not need to run a server. Matter of fact it's
probably overkill. Certain Macintosh server software can run on even a lower 
level machines.(ie MAC HTTP)

With a Windows NT server can you do file sharing with Macs? and can it be setup
hooked up to do phone net?

Changing an installed base of hardware and software is not a wise move,but
networking them so they can all be used is at least a potential better
situation. 

Rick

    A neighbor called me this morning..she's been considering
    Macs and PCs and what to buy.  She has decided to buy a new
    PowerMac 6400/180 and a 15" MultiScan Monitor from SmallDog Electronics.
    $1249 for the PowerMac (or is it a Performa?) and $329 for the monitor.

    These prices beat everything around here.  The 6400/180 comes with
    180MHz 603e, 16MB Ram, 1.6 GB Hard Drive, 8X CD, 28.8 modem, lot
    of software.  For $80 they'll add another 16MB Ram.

    I'm in TechnoLust bad.
    bob

    Bob, you did, of course, advise your neighbor to put up the extra $80
    for 16MB more memory.
    
    Wow!! PowerMac 6400/180, 15" monitor, 1.6GB hard drive, 32MB memory, 8X
    CD and 28.8Kb modem, for under $1,700.  And a bunch of "free" software.
    
    A great deal, indeed.
    
    /Wayne

>    Bob, you did, of course, advise your neighbor to put up the extra $80
>    for 16MB more memory.
    
    I sure did!  And she is.  I really appreciate these guys not gouging
    the customer for extra memory.  Maybe I could buy it for a little
    less...and then I'ld have to put it in for her.  Great deal.

>    Wow!! PowerMac 6400/180, 15" monitor, 1.6GB hard drive, 32MB memory, 8X
>    CD and 28.8Kb modem, for under $1,700.  And a bunch of "free" software.

    I wrote to SmallDog about how they can sell this EXACT machine for
    $1249 when its $1699 at Staples (and this is their 'special' price)...
    He replied....

From:	56405::"Don@smalldoggy.com" "Don Mayer" 31-JAN-1997 12:00:37.83

Thanks for the note, wwe really appreciate the feedback.

We purchased these 6400s at a special price as Apple was stuck with quite 
a few extras at the end of the year (it was the primary reason for last 
quarter's Apple loss) and are passing the savings on to the customers.   
While Staples and most retail stores have the same product at $500 more 
many mail order companies have similar prices to ours as Apple released 
thousands of these machines.

Don

Don Mayer
Small Dog Electronics
High Technology for Low Prices
802-496-7171
802-496-6257 -fax
Don@smalldoggy.com
For complete pricing and inventory status -> www.smalldoggy.com

    Perspective on the deal:  For $1,700 at Staples you get the PowerMac
    6400/180 w/16MB RAM, 1.6GB hard drive, 8X CD and 28.8Kb modem.  That
    ain't bad!
    
    For the same price from SmallDog, you get the same machine with "free"
    15" monitor and 16MB more RAM.  That's real good!
    
    Patience is a virtue!  But who would have guessed?

Before Christmas the 6400/180 was $1899 firm no matter where you went. Right
about Christmas Sears dropped it to $1599 'on sale until 12/27'.  That's
when I bought mine (awesome system).  At MacWorld I saw them for $1239
while the others (Staples etc.) have now dropped to the $1599 - $1699 range
and Small Dogg continues the MacWorld prices...

I love my 6400 (I added 32Mb for a total of 48 and a 17" Sony 200sx monitor
then I hooked up my Zip drive and LaCie scanner...Having fun...I wanna go
home and play some more!

I just have to ignore the lower prices...I knew when I bought it wouldn't 
be long before I saw same or better for cheaper...I just didn't think it
would be THAT MUCH cheaper!


John

>                     <<< Note 875.35 by DPE1::ARMSTRONG >>>
>
>    Can anyone point me to some benchmarks or other comparisons?
>
>    PowerPC 604 versys 601/603?

    I posted this a few back....ask and yee shall receive...

    This came through Evangelista today, unasked...

Subject: PowerPC vs Pentium Benchmarks
Message-ID: <970203180159_-1744319096@emout11.mail.aol.com>

This tidbit is from:

Larry Yaeger, <larryy@apple.com>

Another documented win for PowerMac over Pentium, even with Pro and MMX 
technologies, from an unbiased source, Byte magazine:

<http://www.byte.com/art/9702/img/027bita2.htm>

The URL above has a nice graphic that tells the story quickly.

The story is that Byte benchmarkers ran a 200 MHz 604e PowerPC Mac 
against 200 MHz Pentium, Pentium Pro, and Pentium with MMX systems, plus 
one Cyrix 166+ system. They tested five basic functions in Photoshop: 
Arbitrary Rotate, Unsharp Mask (default), Unsharp Mask (custom), Gaussian 
Blur, and RGB to CMYK.

On three of the tests--Arbitrary Rotate, Unsharp Mark (custom), and RGB 
to CMYK--PowerMac blew away Pentium+MMX by about a factor of two! On the 
other two tests, PowerMac was within a few percent of Pentium+MMX 
performance. PowerMac soundly beat all other Pentium products (non-MMX) 
on all five tests (often by factors of three, four, and even more).

All MHz are NOT created equal. And PowerMacs rule!  

    
    One caveat, Photoshop is still really a Mac Application. Adobe has
    already said that they will improve Photoshop's speed on other
    platforms. They said it is written in a development environment (MacApp
    I think) that is a lot more Mac freindly and the ported code is not
    tuned for other processors. Eventually they will tune the code for
    other processors and the 2:1 ratio advantage that the Mac enjoys today
    will go away. 
    
    Mario

In the Evangelist posting yesterday, there was another
benchmark, timings of Mathematica on various platforms...

I have not looked at the URL referenced, but the gist of
the benchmarks seem to be...

a 120MHz 7600 (604) is about the same speed as a
  200MHz Pentium Pro running NT (surprises me)

PowerMac Clones rule the performance world

a Pentium Pro running NextStep is about twice as fast as the same
    hardware running WindowsNT....could this be good news for Rhapsody?


Date: Tue, 11 Feb 1997 20:02:21 -0500 (EST)
From: MacWay@aol.com
Subject: Mathematica Benchmarks
Message-ID: <970211200033_-1776881752@emout03.mail.aol.com>

Keyword: Market by market, Science

This tidbit is from:

Larry Yaeger, <larryy@apple.com>

Check out these Mathematica benchmarks: 

<http://fampm201.tu-graz.ac.at/karl/timings30.html>

It's a set of benchmark timings of Mathematica on various platforms. The 
datapoints are a little difficult to correlate, since RAM sizes and L2 
cache sizes vary a lot, and I don't know how memory intensive this 
particular evaluation is. But there are a few interesting observations to 
be made from these numbers.

[All benchmark performance numbers are given relative to a reference of 
1.0 for a PowerMac 7600/120 (a 120 MHz 604 processor), and higher is 
better.]

First, a PowerMac clone is at the absolute top of the list with a 
benchmark performance value of 2.09. That's nice.

Second, four of the top five machines are PowerMacs, and the only 
Intel-based machine in those top five is running NeXTStep v3.3.

Third, following up on that last observation, the tests happen to include 
benchmarks for four different 200 MHz Pentium Pro systems--one running 
NeXTStep v3.3, two running WindowsNT v4.0, and one running Windows 
95--which yield the following benchmark performance numbers:

NeXTStep v3.3 1.86
Windows NT v4.0 1.071
Windows NT v4.0 1.01
Windows 95 0.956

So with approximately the same hardware (certainly the same CPU), 
NeXTStep outperforms both Windows NT and Windows 95 by nearly a factor of 
two. This bodes very well for our future MacOS plans!  

    In the System7 note I asked about MacOS 7.6, and whether anyone
    has tried it yet...wondering about the claimed 20% speedup over 7.5
    due to additional native OS code.  Still wondering about that.  And
    any other review...
    
    I notice that in the new Performa6400 I just got, it comes at OT1.1.0.
    Is there any problem doing the OT 1.1.2 upgrade?  I wish there
    were some easy place to find all the latest versions of all the
    various bits in these machines so I could compare them with what
    I have.  Or maybe its not that important to keep up?  I'm wondering about
    upgrading to 7.5.5 with the interrupt extension....??

    Today I ran Speedometer against the 6400/180.  It comes to about an
    average of 5.285 (Quadra 605 is 1).  but it varies a lot (as you would
    expect).

    cpu - 7.825
    graphics - 3.3446
    disk - 4.150
    Math - 464.223

    For comparison, my LCIII rated about .5!!!!

    One more question....I'm wondering if there is some way to verify
    that I do have the 256K cache in the machine.  Some things mentioned in
    TechTool that I dont recognize are

    VIA1
    SCC

    These are listed in the 'hardware' report.  TechTool makes excuses
    about being unable to truly recognize many modern Macs, especially
    performas, and it recognizes this machine as a UMAX PowerMac.

    One other question....TechTool claims this machine has a 40MHz Bus,
    but GURU claims 45MHz???

    So far this feels like a great machine.
    bob

    
    ! One more question....I'm wondering if there is some way to verify
    ! that I do have the 256K cache in the machine.  Some things mentioned in
    ! TechTool that I dont recognize are

    Bob;
    
    I believe the application called Cache-22 will give you this info. Just
    wish I could remeber where I got it from. Maybe, Newer Co. The version
    I have is v1.3b2. If you can't find it let me know and I'll put it up
    on my PW/Mac server.
    
    Regards;
    Dave Y.

>    I believe the application called Cache-22 will give you this info. Just
>    wish I could remeber where I got it from. Maybe, Newer Co. The version
>    I have is v1.3b2. If you can't find it let me know and I'll put it up
>    on my PW/Mac server.

    Ah...thanks.  I just looked in the archive and found

    L2_CACHE_CHECK.INSTALLER

    and it did exactly what I wanted....reported that I have a properly
    functioning level 2 cache installed.
    bob