Performance of student version of Mathematica?
Kale Beckwitt
VERY good deal on the student version of the software through the store at
the student store at my university. Unfortunately, this version does not
utilize a math-co-processor. I was wondering if anyone has any experience
with this version expecially regarding how this limitation effects software
performance with large data sets.
Thanks in advance!
Kale Beckwitt
ka...@uclink2.berkeley.edu
/----------------------------------------------------------\
| ka...@uclink2.berkeley.edu |
| Kale Beckwitt I suppose if we couldn't |
| Clark Kerr Campus laugh at things that don't |
| 2601 Warring Street make sense, we couldn't |
| Box #364 react to a lot of life. |
| Berkeley, CA 94720 -- Hobbes |
| (510)664-3496 |
\----------------------------------------------------------/
John Burnette
Kale Beckwitt <ka...@uclink2.berkeley.edu> wrote:
>I am going to purchase the Windows version of Mathematica and I can get a
>VERY good deal on the student version of the software through the store at
>the student store at my university. Unfortunately, this version does not
>utilize a math-co-processor. I was wondering if anyone has any experience
I suppose it is a matter of opinion whether to capitalize the "VERY" in
your post. I personally think it is absurd to force the non-use of an FPU
for MATHEMATICS software!! As we move into faster processor chips
(particularly the powerpc) will we get pushed into the absurd situation of
being forced to run the software in emulation mode? The NeXT academic
community used to get the software INCLUDED with their computers however
when they paid for "upgrades" suddenly their multiuser programs (the
point of UNIX computing) suddenly had code inserted to force it to be
single user.
Deep breath.... I guess my point is I have severe philosophic problems
with a company who intentionally cripples their software before selling
it to educational organizations. In my opinion if Wolfram wants to sell
less for less to education, I would BEG them to consider selling full
speed software with smaller libraries. I'd say most, by far, of the
academic users of Mathematica could get away with 25% of the capability
of the program. I personally believe they would get far more market
penetration in this manner.
Back to your particular question, after using the PowerPC native version,
there is NO WAY I would go back to a non-FPU version of any kind. It has
to be 5-10 times faster.
just my $.02
James Moore
: I am going to purchase the Windows version of Mathematica and I can get a
: VERY good deal on the student version of the software through the store at
: the student store at my university. Unfortunately, this version does not
: utilize a math-co-processor. I was wondering if anyone has any experience
: performance with large data sets.
: Thanks in advance!
My understanding is that if you purchase the student version, you can
upgrade to the full-blown version for a fee. The total cost is still
less than buying the full-blown version outright.
: Kale Beckwitt
: ka...@uclink2.berkeley.edu
: /----------------------------------------------------------\
: | ka...@uclink2.berkeley.edu |
: | Kale Beckwitt I suppose if we couldn't |
: | Clark Kerr Campus laugh at things that don't |
: | 2601 Warring Street make sense, we couldn't |
: | Box #364 react to a lot of life. |
: | Berkeley, CA 94720 -- Hobbes |
: | (510)664-3496 |
: \----------------------------------------------------------/
--
Jim
moo...@assets.wharton.upenn.edu
(215) 898-3589
Brian D. Cowell
: In article <3m55ub$g...@news0.cybernetics.net>,
: Kale Beckwitt <ka...@uclink2.berkeley.edu> wrote:
: >I am going to purchase the Windows version of Mathematica and I can get a
: >VERY good deal on the student version of the software through the store at
: >the student store at my university. Unfortunately, this version does not
: >utilize a math-co-processor. I was wondering if anyone has any experience
:
: for MATHEMATICS software!! As we move into faster processor chips
Just the humble opinion of a Frosh Boring Calc Student:
I paid ~$100 for it. Dunno what the normal cost is.
It isn't horribly slower than what I use in our lab on Sparc 5's.
Then again, I only use it for simple stuff. As for huge operations,
mine didn't crash until I had ~20 megs of (kernel + notebooks) open
on a system with 8 meg ram & 20 meg swap file. OK, so this might
be completely irrelevant to your question, I don't know.
As for the FPU thing, they probably just want you to upgrade to
the Enhanced version.
Anyway, if you send me some simple code to run, I'll time it on
my student version if it helps ya.
Brian
bdco...@mtu.edu
Roger Frederi Clark
in...@wri.com and here's what they said. The student version will not use
a math coprocessor, even if present. THe rep said that this makes
little difference for normal calculations, but for large data
sets, numerical calculations, or intensive graphics work, it translates into
about 4 times slower. Unfortunately, this is what I use Mathematica for,
so I guess I'll have to go with the full version.
Roger Clark
--
| /\ Biker, SCUBA diver, philosopher at heart.
hel...@strauss.udel.edu |/^vv^\ o. "We can only observe and theorize,
University of Delaware | \ _>~<\_ we can never know." -A. Einstien
Chemistry (and others) | \__(*) '(*)__________________________________
Demetri Mouratis
|> In article <3m55ub$g...@news0.cybernetics.net>,
|> Kale Beckwitt <ka...@uclink2.berkeley.edu> wrote:
|> >I am going to purchase the Windows version of Mathematica and I can get a
|> >VERY good deal on the student version of the software through the store at
|> >the student store at my university. Unfortunately, this version does not
|> >utilize a math-co-processor. I was wondering if anyone has any experience
|>
|> I suppose it is a matter of opinion whether to capitalize the "VERY" in
|> your post. I personally think it is absurd to force the non-use of an FPU
|> for MATHEMATICS software!! As we move into faster processor chips
|> (particularly the powerpc) will we get pushed into the absurd situation of
|> being forced to run the software in emulation mode? The NeXT academic
|> community used to get the software INCLUDED with their computers however
|> when they paid for "upgrades" suddenly their multiuser programs (the
|> point of UNIX computing) suddenly had code inserted to force it to be
|> single user.
|>
|> Deep breath.... I guess my point is I have severe philosophic problems
|> with a company who intentionally cripples their software before selling
|> it to educational organizations. In my opinion if Wolfram wants to sell
|> less for less to education, I would BEG them to consider selling full
|> speed software with smaller libraries. I'd say most, by far, of the
|> academic users of Mathematica could get away with 25% of the capability
|> of the program. I personally believe they would get far more market
|> penetration in this manner.
|>
|> Back to your particular question, after using the PowerPC native version,
|> there is NO WAY I would go back to a non-FPU version of any kind. It has
|> to be 5-10 times faster.
|>
|> just my $.02
|>
|>
|>
|>
I have to say, your response to Kale's question has some problems. Most of them are in the second paragraph.
To start, the Student Version of Mathematica is not what is sold to "educational orgainizations". In fact, universities typically own an enhanced versions site license for a mix of platforms. This is the case at the University of Illinois. The Student Version, as its name implies, is marketed towards the student.
To claim that Wolfram Research has "intentionally crippled" the student version is absurd. The decision was to offer a lest costly product so that students, typically college ones, could afford the software. The lower price means lower performance. In this case, lack of coprocessor support.
"I'd say most, by far, of the academic users of Mathematica could get away with 25% of the capability" Again, I disagree. I, as a student, find myself using 50-60% of the packages on a regular basis. I have probably loaded closer to 80% of the packages, but rarely used some of them. To reduce the number of "libraries" doesn't make sense since they are readily available to any Mathematica user.
You were right in one respect. The enhanced version is faster. 5-10 times faster, I don't think so. 4 times, in intesive calculations, probably. The student version will give the same results as any other version. Is it worth it to buy the student version vs the enhanced? I don' know. How impatient are you and how much do you have to spend?
Your PowerPC is a great machine. Try running the Student Version on it, you'll be surprised by the results.
John Burnette
Roger Frederi Clark <hel...@strauss.udel.edu> wrote:
>I was curious about this possibility also, so I contacted the
>in...@wri.com and here's what they said. The student version will not use
>a math coprocessor, even if present. THe rep said that this makes
>little difference for normal calculations, but for large data
>sets, numerical calculations, or intensive graphics work, it translates into
>about 4 times slower. Unfortunately, this is what I use Mathematica for,
etc etc...
Hello? Gee, unless you want to do something UNUSUAL like NUMERICAL
CALCULATIONS...(And what kind of mathematica user would have that in
mind....) And compare the native code version of mathematica running on a
powerpc macintosh to a student version running under 68040 emulation on
the same computer. That is NOT a 4 times slower. Wolfram's OWN
ADVERTISING gives numbers like 17 seconds versus 109 seconds (on a Quadra
700, presumably running an enhanced version...) For those with a new
PowerMac check out the sample software from Wolfram on the CD sampler.
Questions for WOLFRAM rep...(surly mode on...)
1) Will there ever be a native (student) version of Mathematica for the
powerpc. Hypothetically, if there was no MacIntosh for sale powered
by something other than a power-pc would this be a factor in your
policy, or should we just be thankful that Mathematica wasn't first
created in the 8088 era?
2) What will happen to the current user base when OS for Mac and INTEL
computers become multi-tasking? (Both mac and windows claim that is
coming within the next year.) Will you slip code into student version
which checks to see if there is a session already running? Is that code
already there?
As an aside, I have been using Mathematica while teaching my high school
math (and computing) courses since 1988-89. I am the leading proponent of
increasing computer use, specifically Mathematica, at my current school -
Choate Rosemary Hall. While visiting Middlesex academy for a school
recertification visit last week, I noticed that Mathematica was installed
but virtually unused on their computers. (Note of interest, Middlesex and
Choate have both been listed in ads from Wolfram..)
I am becoming the math department chair at a third school, Kinkaid in
Houston Texas. I would very much like to continue working with
Mathematica there, but there is no way I will pay more than $200 per
license and, on principle, no way I would purchase any mathematics
software which has intentionally been crippled so as not to make use of
either an advanced processor or an FPU. Honestly, the only way I would
recommend purchasing Mathematica would be to purchase used NeXT
workstations which include an academic license.
Wolfram must realize that they have almost ZERO penetration of the
(fairly) large high school market. I firmly believe that their bottom
line would be increased significantly if they came out with a
"Mathematica lite" version which used a kernal which actually fits onto a
8 meg computer. Provide the full processing speed and let the schools
purchase more libraries as they decide they need them.
As for me, MathCad is offered for $49 with the Maple libraries. I don't
want to learn a new environment, but I suppose it isn't out of the
question either. It's too bad. I've carried a torch for Mathematica for a
very long time.
Gerhard Wesp
Kale Beckwitt <ka...@uclink2.berkeley.edu> wrote:
>I am going to purchase the Windows version of Mathematica and I can get a
>VERY good deal on the student version of the software through the store at
>the student store at my university. Unfortunately, this version does not
>utilize a math-co-processor. I was wondering if anyone has any experience
>performance with large data sets.
>
I wouldn't recommend using Mma for *really* large data sets,
unless I really had too. Mathematica is quite powerful in doing symbolic
math, but your problem is likely to be of numerical nature. Mma's data
structures are highly dynamic, this gives it a lot of flexibility, but
imposes also some performance limitations.
However, if you do numerical calculations, you might experience
a performance degradation of up to 90%, especially if it "doesn't utilize
a math-co-processor", as you say.
If you're doing linear algebra, have a look at octave, which is a frontend
to LAPACK. It will be much faster than *any* symbolic math
program. However, if you're doing symbmath only, the presence of a FPU
shouldn't make any difference.
Hope this helps!
-Gerhard
John Burnette
Demetri Mouratis <dm5...@glhpx15.cen.uiuc.edu> wrote:
>I have to say, your response to Kale's question has some problems. Most of
them are in the second paragraph.
>
>To start, the Student Version of Mathematica is not what is sold to
>"educational orgainizations". In fact, universities typically own an
>enhanced versions site license for a mix of platforms. This is the
>case at the University of Illinois. The Student
>Version, as its name implies, is marketed towards the student.
>
I agree. The Student version is sold THROUGH educational organizations.
>To claim that Wolfram Research has "intentionally crippled" the student
>version is absurd. The decision was to offer a lest costly product so
>that students, typically college ones, could afford the software. The
>lower price means lower performance. In this case, lack of coprocessor
>support.
My choice of language was intentionally confrontational, but by no means
absurd. Point in fact, Wolfram REMOVED support for FPU for the
educational version. Secondly, they inserted code into UNIX licenses so
that they could restrict the number of concurrent sessions. Both were
intentional EXTRA work which served only to artificially reduce the value
and usefulness of academic licenses. I believe my labelling of this
effort as an "intentional crippling" of their software is extremely
accurate. Suppose Wolfram had taken the approach that the student version
would use the exact software as the full version but a timing loop had
been added in order to quadruple the calculation time. Logically what is
the difference between such and approach and the one Wolfram took?
>
>"I'd say most, by far, of the academic users of Mathematica could get away
>with 25% of the capability" Again, I disagree. I, as a student, find
>myself using 50-60% of the packages on a regular basis.
Perhaps I have been caught in a sweeping generality. Let me try again.
The statement above would be more accurate if it refered to beginning math
students working primarily with Calculus. It is certainly absolutely
on target for the high school environment. The source of my frustration is
my increasing certainty that there is little chance I will be able to
convince my new school to fund the purchase of mathematica.
>
>You were right in one respect.
(Only..) one?? How kind of you to say...
>The enhanced version is faster. 5-10 times faster, I don't think so.
>4 times, in intesive calculations, probably. The student version will give
>the same results as any other version. Is it worth it to buy the student
>version vs the enhanced? I don' know. How impatient are you and how
>much do you have to spend?
In truth this is one area where I was conservative. By Wolfram's own
advertising, it can be 10 times faster than a Quadra 700 (presumably
running the enhanced version. Whether it is "worth" purchasing a student
version is indeed related to how impatient I am, as well as how much I
have to spend - but it also is related to how much I am willing to be
treated in a patronizing fashion and what the capabilities and prices of
other packages are.
>
>Your PowerPC is a great machine. Try running the Student Version on
>it, you'll be surprised by the results.
Oh I was indeed surprised by the results of running the Student Version
2.2 in emulation mode. It hung the computer. (For your information, my
computer is a NeXTstation which included a MULTIUSER version of Mathematica.)
Thanks for chatting. (BTW, are you an employee of Wolfram?) I believe I've
raised several pointed, direct questions. Perhaps I've been too strident
for some of you. I'd like to hear someone who will identify themselves as
a Wolfram representative answer the direct questions I posed.
Gary L. Gray
Burnette) wrote:
[stuff deleted]
In speaking with some people at Wolfram, it is my understanding that in
the next version of Mma (2.3 or 3.0 or whatever), the student version will
not be crippled in any way. If I understood them correctly, students will
have one month to send Wolfram in a copy of their student ID and they will
then be sent a code that will prevent Mma from "crippling" (i.e., removing
fpu/PPC support) itself after one month. This sounds great. If it is true,
I am really glad they are doing this.
Can someone at Wolfram confirm, deny, and/or elaborate on this?
Thank you,
--
Gary L. Gray | (814) 863-1778
Assistant Professor | (814) 863-7967 (fax)
Engineering Science & Mechanics | glg...@engr.psu.edu
Penn State University | AOL: GLGray
Ian Collier
Burnette) wrote:
> My choice of language was intentionally confrontational, but by no means
> absurd. Point in fact, Wolfram REMOVED support for FPU for the
> educational version. Secondly, they inserted code into UNIX licenses so
> that they could restrict the number of concurrent sessions. Both were
> intentional EXTRA work which served only to artificially reduce the value
> and usefulness of academic licenses. I believe my labelling of this
> effort as an "intentional crippling" of their software is extremely
> accurate. Suppose Wolfram had taken the approach that the student version
> would use the exact software as the full version but a timing loop had
> been added in order to quadruple the calculation time. Logically what is
> the difference between such and approach and the one Wolfram took?
> >
<<Stuff deleted>>
> Thanks for chatting. (BTW, are you an employee of Wolfram?) I believe I've
> raised several pointed, direct questions. Perhaps I've been too strident
> for some of you. I'd like to hear someone who will identify themselves as
> a Wolfram representative answer the direct questions I posed.
Wolfram Research has not removed performance from any educational
versions of Mathematica.
The Macintosh and PC versions of Mathematica have always been
available in both Standard and Enhanced versions. The Standard
version is designed to run on computers that do not support
hardware floating point instructions. The binaries in Mathematica
for Students for Macintosh and Windows are identical to the
respective Standard versions. The only difference between
Mathematica for Students and the Standard professional version
for either platform is in documentation and the way that technical
support is provided.
What Wolfram Research has in fact done is to take an existing
product and make it available at a much lower price to full
time students. Any extra work went into creating more accessible
documentation.
Note that the full Enhanced version of Mathematica is available
to educational institutions, faculty and students at academic
discount prices.
I hope this answers your questions.
Best regards,
--Ian
----------------------------------------------------------------
Ian Collier
Technical Sales Support
Wolfram Research, Inc.
----------------------------------------------------------------
Tel (217) 398-0700 Fax (217) 398-0747 ia...@wri.com
Wolfram Research Home Page http://www.wri.com/
----------------------------------------------------------------
Gerhard Wesp
Demetri Mouratis <dm5...@glhpx15.cen.uiuc.edu> wrote:
>To claim that Wolfram Research has "intentionally crippled" the student
>version is absurd. The decision was to offer a lest costly product so
>that students, typically college ones, could afford the software. The
>lower price means lower performance. In this case,
>lack of coprocessor support.
I don't think the claim is too absurd. Well, "crippled" is a strong word, but
it describes the situation quite well in this case.
I mean, as compared to the normal version "with coprocessor support",
Wolfram *adds* some pieces of code that makes the software slower.
You agree on that?
Since Wolfram doesn't release the sources, we don't know what kind of
code they add, we only see it makes the software slower, so we might assume
it is delay loops. Delay loops are the standard way of making software
slower. Period.
The point is, they have to put some effort in making the software slower.
So, the *additional* effort Wolfram have to invest makes their costs lower!?
Well, the development of delay loops isn't too difficult, but perhaps
they have very "sophisticated" delay loops, like floating point emulation?
We simply don't know.
Greetings,
-Gerhard
P.S.: PLEASE trim your lines to the standard 78 columns.
Rob Anomaly Boy Keller
> to educational institutions, faculty and students at academic
> discount prices.
>
> I hope this answers your questions.
>
> Best regards,
>
> --Ian
>
> ----------------------------------------------------------------
> Ian Collier
> Technical Sales Support
> Wolfram Research, Inc.
> ----------------------------------------------------------------
> Tel (217) 398-0700 Fax (217) 398-0747 ia...@wri.com
> Wolfram Research Home Page http://www.wri.com/
> ----------------------------------------------------------------
In fact I bought the Enhanced version at academic pricing. I am a senior
year university math student, and I recently bought a PowerMac. I made
sure that when I bought my powermac to budget for enough ram, an L2 cache,
a great monitor, and plenty of drive space. What good is that PPC chip if
you can't exploit it? Anyway, with that attitude, I also decided to buy
the Enhanced version of mathematica. Admittedly it was expensive at $350
for a starving student, but at the same time, I wanted to take advantage
of a decent computer system. So keep in mind, that if you spent over
$3000 to put together a decent computer system, $350 is not that much.
$350 seemed like a lot to me, until I realized that the machine and
peripherals cost me $6000. Just set your priorities. Also, I have no
money, just a nice parent willing to cosign a low interest loan.
rob
ps - it was worth the extra money for the peace of mind that the better
tech support provides.
--
"I turn away in fright and horror from the lamentable evil
of functions which do not have derivatives"
- Charles Hermite