From: mark@nwra.com (Mark Baldwin) Subject: Benchmarks for IDL for Windows/Windows NT? Message-ID: Date: Fri, 3 Mar 1995 21:22:11 GMT I am considering using IDL on a Windows NT system as an alternative to my current Sparc 10. If you are running the Windows or Windows NT version of IDL (on any CPU) would you please execute "time_test" and let me know the total time and your system configuration? I will post a brief summary of the results. Thanks, Mark Baldwin mark@nwra.com -------------------------------------------------- From: mark@nwra.com (Mark Baldwin) Subject: Summary: "time_test" for IDL on Windows systems Message-ID: Date: Thu, 16 Mar 1995 19:39:29 GMT The following is a summary of the responses I received in response to my post concerning the IDL benchmark "time_test" on Windows and NT systems: System: Gateway 90MHz Pentium, 32Mb RAM, Windows for workgroups. IDL 3.6.1: 15.59 seconds 1st Beta IDL 4.0: 11.23 seconds 2nd Beta IDL 4.0: ~16.8 seconds System: 486-DX2 66MHz 8Mb RAM IDL 3.6.1: 47.74 seconds. An error occurred in test #15, resulting in 0.0 seconds for that test. System: 90MHz Pentium, 32Mb RAM, Windows for Workgroups 3.11 IDL 3.6.1: 16.4 - 16.6 seconds. Same system except Gateway 90MHz w/ very fast disk: 14.2 seconds System: 486/33MHz Windows 3.1 IDL 3.1.2: 73.71 seconds. ----------------------------------------------------------------- For comparison, my Sparc 10/41 takes 21 seconds. The Pentium systems are faster, but there are a few points to be cautious about. Disk speed becomes relatively important with the fast systems. Also, you will need to run time_test a few times because it gets faster after the first try. The big question concerns the two beta versions of IDL 4.0. The first beta is much faster than 3.6.1, while the second beta is actually slower. Let's hope RSI keeps the speed of the first beta in the 4.0 release, which is due in late spring or early summer. Can anyone familiar with these beta versions explain these speed differences? Has anyone run the NT version of IDL on a Pentium system? -Mark Baldwin mark@nwra.com -------------------------------------------------- From: fskmjm@puknet.puk.ac.za (Mathews, MJ) Subject: Re: Summary: "time_test" for IDL on Windows systems Message-ID: Date: Mon, 20 Mar 1995 07:30:06 GMT Summary: "time_test" for IDL on Windows systems In article mark@nwra.com (Mark Baldwin) writes: >The following is a summary of the responses I received in response to >my post concerning the IDL benchmark "time_test" on Windows and NT >systems: I find on my 486DX2 with 8MB RAM, Windows 3.1, IDL ver 3.5.1 that the test is very sensitive to the size of my output window. Does this make sense????? -------------------------------------------------- From: smithjp@sage.cc.purdue.edu (Joseph Smith) Subject: Platform recommendation/tradeoffs? Message-ID: <3mhc42$6lt@mozo.cc.purdue.edu> Date: 12 Apr 1995 20:10:42 GMT Sender: smithjp@sage.cc.purdue.edu (Jospeh Smith) Keywords: IDL Mac Iris I am about to order some flavor of IDL (Mac or Unix) and I'm wondering if anyone can help me gauge just how much performance I loose by getting the Mac product. (This saves me enough money to buy a Q630 cpu) I have a SG Iris Indigo with lots of memory and some disk space that I could be using for IDL, or I could put it on the aforementioned Q630. If intense image processing were among my requirements it would be a pretty easy decision, but the data that I'll be working on is basically about 16 or 32 128x128 matrices - not too big of a problem from a graphic standpoint (I suppose I would rarely need to look at more than 1 128x128 image at a time, for instance) This problem has a lot of number crunching - 128x128 non linear least squares fits of at least 4 parameters per fit. Any comments? Any experience using IDL on a 68040? TIA, Joe Smith -------------------------------------------------- From: bowman@csrp.tamu.edu (Kenneth P. Bowman) Subject: Re: Platform recommendation/tradeoffs? Message-ID: Date: Wed, 12 Apr 1995 21:37:38 -0500 In article <3mhc42$6lt@mozo.cc.purdue.edu>, smithjp@sage.cc.purdue.edu (Joseph Smith) wrote: > I am about to order some flavor of IDL (Mac or Unix) and > I'm wondering if anyone can help me gauge just how much performance > I loose by getting the Mac product. (This saves me enough money > to buy a Q630 cpu) I have been pretty impressed with IDL running on a PowerMac 6100/60, which is the slowest (non-upgraded) PowerMac. That is in comparison with some IBM RS6000 and DEC Alpha workstations. I have not done any timing comparisons though. I'm not sure I would want it on a 68040 though. How about the new PPC 603e based Mac, the 5200 or whatever it's called? List price $1599, I think. Perhaps RSI has some comparison numbers? Regards, Ken Bowman -- Dr. Kenneth P. Bowman 409-862-4060 Associate Professor 409-862-4132 fax Department of Meteorology bowman@csrp.tamu.edu Texas A&M University PP-Glider College Station, TX 77843-3150 -------------------------------------------------- From: daft@debussy.crd.ge.com (Chris Daft) Subject: Re: Platform recommendation/tradeoffs? Message-ID: Date: Fri, 14 Apr 1995 00:41:03 GMT In-Reply-To: bowman@csrp.tamu.edu's message of Wed, 12 Apr 1995 21:37:38 -0500 Sender: usenet@crdnns.crd.ge.com (USENET News System) Nntp-Posting-Host: debussy.crd.ge.com In article bowman@csrp.tamu.edu (Kenneth P. Bowman) writes: Ken Bowman wrote (on comparative IDL performance): Perhaps RSI has some comparison numbers? Regards, Ken Bowman I second the request for some comparison of IDL's performance on various CPUs. I have been using IDL on Suns for a long time, but may be required to switch to a Pentium machine for some period, and I'd like to know what the performance differences would be. I wonder if some kind of standard IDL test suite could be developed, kinda like a SPECmark? I think this might actually be of competitive advantage to RSI. A colleague recently ported some IDL code [ultrasound transducer simulations] to Matlab and found it to be *twice* as slow. . . -- Chris M.W. Daft KWC-1336, GE CR&D 518-387-6615 P.O. Box 8 518-387-7512 (fax) Schenectady, N.Y. 12301-0008 -------------------------------------------------- From: ebstein@lexitek.com (Steve Ebstein) Subject: Re: Platform recommendation/tradeoffs? Message-ID: Date: Thu, 13 Apr 1995 09:26:02 -0500 In article , bowman@csrp.tamu.edu (Kenneth P. Bowman) wrote: > I have been pretty impressed with IDL running on a PowerMac 6100/60, which > is the slowest (non-upgraded) PowerMac. That is in comparison with some > IBM RS6000 and DEC Alpha workstations. I have not done any timing > comparisons though. I'm not sure I would want it on a 68040 though. How > about the new PPC 603e based Mac, the 5200 or whatever it's called? List > price $1599, I think. The PowerMac version is easily 4X faster than a 68040 version (7100-66MHz vs 33MHz 68040). -- Steve Ebstein ebstein@lexitek.com -------------------------------------------------- From: kennealy@mv.mv.com (Jack Kennealy) Subject: Re: Platform recommendation/tradeoffs? Message-ID: Date: Fri, 14 Apr 1995 04:04:20 GMT Nntp-Posting-Host: mv.mv.com Sender: usenet@mv.mv.com (System Administrator) daft@debussy.crd.ge.com (Chris Daft) writes: >I second the request for some comparison of IDL's performance on >various CPUs. I have been using IDL on Suns for a long time, but may >be required to switch to a Pentium machine for some period, and I'd >like to know what the performance differences would be. >I wonder if some kind of standard IDL test suite could be developed, >kinda like a SPECmark? Every IDL distribution includes a copy of time_test.pro, which is exactly the standard test suite you refer to. I've run this on quite a few different machines, but I don't have the data here right now in my home office. However, I do recall that my 90MHz Pentium system fares somewhat better than a a SUN SPARCstation-10 on this test. I was a little surprised by that result, so I used one of my own Fortran FFT routines to compare my Pentium with the SPARC-10 and got pretty much the same result. The Pentium is indeed impressive. -- =============================================================== Dr. Jack Kennealy, Nashua, NH kennealy@mv.mv.com =============================================================== -------------------------------------------------- From: foresto@sun16.NoSubdomain.NoDomain (Vincent Coude' du Foresto) Subject: Re: Platform recommendation/tradeoffs? Message-ID: <3mlqtg$4ih@sun0.urz.uni-heidelberg.de> Date: 14 Apr 1995 12:47:43 GMT Sender: foresto@sun16 (Vincent Coude' du Foresto) Keywords: IDL Mac Iris In article <3mhc42$6lt@mozo.cc.purdue.edu>, smithjp@sage.cc.purdue.edu (Joseph Smith) writes: |> I am about to order some flavor of IDL (Mac or Unix) and |> I'm wondering if anyone can help me gauge just how much performance |> I loose by getting the Mac product. (This saves me enough money |> to buy a Q630 cpu) |> |> I have a SG Iris Indigo with lots of memory and some disk space |> that I could be using for IDL, or I could put it on the aforementioned |> Q630. If intense image processing were among my requirements it would |> be a pretty easy decision, but the data that I'll be working on is |> basically about 16 or 32 128x128 matrices - not too big of a problem |> from a graphic standpoint (I suppose I would rarely need to look at |> more than 1 128x128 image at a time, for instance) |> |> This problem has a lot of number crunching - 128x128 non linear |> least squares fits of at least 4 parameters per fit. |> |> Any comments? Any experience using IDL on a 68040? |> I ran a small testbench between a SPARC20, a PowerMac and a Quadra. The test included computing a Fast Fourier Transform on a 1024x1024 float array, and 1000 times computing a 1024 FFT. Here are the results: Machine FFT 1024x1024 1000 FFT 1024 Sparc 20 8.5s 3.4s PowerMac 7100/66 14.6s 7.4s Centris (68040@25MHz + FPU) 39.8s 33.4s A few comments: 1- Take it for what it's worth. It's only a **vague** indicator of raw processing power. I does not tell anything about disk I/O, where the Unix machine outperforms the Mac several times. The Mac OS file manager is notoriously slow (it is still emulated), but this should improve in the future version of the system to be delivered in 1 year (does anyone have precisions or more quantitative information on this?) 2- The Mac and PowerMac are running demo versions of IDL (which refuses to run the built-in testbench (run in 39s by the Sparc20): I don't know why this feature has been inhibited on demo versions btw). The PowerMac is running "native" code, however it's not that much faster than the Centris. I suspect most of the "native" code is actually still emulated. On FPU intensive operations a PowerMac running real native code is usually about 8 times faster than a 25MHz Quadra. All machines here were running IDL 3.6 with sufficient memory (>25Mb) for the test. 3- Notice how the ratio of the times needed to perform the two test benches changes with the processor. Does anybody have a clue? Also, I should say that the "look and feel" of IDL for the Mac is surprising to say the least: buttons do not have the standard Mac appearance, windows are most often not resizable, etc... Clearly IDL is not a Mac application, it's an application from the mainframe world that has been ported to the Mac, with minimal changes. On a 32768 color monitor, I was not able to get allocate 256 colors for IDL while keeping 3D-looking system windows in the background. One can only hope that most of these non-Mac quirks will disappear with version 4.0. Does anyone know how IDL 4.0 for the Mac will look like? Will it be any faster than 3.6 on PowerPC machines? Another Mac-related question: does anybody know if it's possible to produce "real color" pictures (from green-red-blue frames for example) on a Mac with a 12 bit (or more) color monitor? My conclusion: if you are doing only data crunching, use few widget applications (which look ugly on the Mac, what an irony!), and do not want to spend too much money, then probably your best deal would be to buy a PowerPC Mac 6100/66. I would strongly discourage buying a Quadra, even if that means saving a few hundred bucks. IDL and the MacOS are going to run faster on the Power Macs as more and more of the code runs native, while it might be very well possible that RSI will discontinue IDL for 680x0 CPUs in a few years. If your application has intensive file I/O, then you'd better stay with a Unix machine. To give an order of magnitude, loading a 256x256x10 data cube is unpractically long on the Mac for the moment (maybe 1min or more, against about 10-15s on the Sparc20). Vincent Foresto Max-Planck Institut fuer Astronomie, Heidelberg -------------------------------------------------- From: Russ Welti Subject: Re: Platform recommendation/tradeoffs? TIME_TEST Message-ID: Date: Fri, 14 Apr 1995 08:50:40 -0700 X-Sender: rwelti@chroma In-Reply-To: There is an IDL routine, called TIME_TEST, which I have used to compare IDL running on 6 different (Unix) CPUs here in my department, and it was very convenient. I made a color plot of the results, which showed that my machine was one of the slowest ones, and will use that to justify a CPU upgrade to my SPARC LX...! [ You should have seen the DEC Alpha OSF ] It is easy to do if you already have access to one of the machines which you are curious about. I agree that a compounded list of many different machines all together would be a nice reference. P.S. Anyone running or considering running IDL on a PowerMac should go get the free extension for speeding up floating point operations. It made a 20% improvement in our Mac's speed running most scientific applications. The Mac is now faster than my SPARC on many operations. Thanks go to Fred: > You might get a higher speed on the PowerMac if you would install > a new floating point math library (made by Apple). It's called MathLib. > Ask or read about it in a comp.sys.mac.* newsgroup. (I don't know the > details as I don't have a PowerMac. The only thing I do know is that > the MathLib in ROM is ``slow'' when it comes to transcendental functions... > The software patch, i.e., the MathLib extension, solves the speed problem.) > > Good luck, > Fred. (walsteyn@fys.ruu.nl) / Russ Welti /-\ (c-g) University of Washington \-/ Molecular Biotechnology / PO Box 352145 /-\ Seattle, WA 98195 (a-t) rwelti@u.washington.edu \-/ (206) 685 3840 voice (206) 685 7344 FAX / http://chroma.mbt.washington.edu/graphics/gif/russ.gif -------------------------------------------------- From: kennealy@mv.mv.com (Jack Kennealy) Subject: Re: Platform recommendation/tradeoffs? Message-ID: Date: Sat, 15 Apr 1995 03:13:05 GMT Keywords: IDL Mac Iris Nntp-Posting-Host: mv.mv.com Sender: usenet@mv.mv.com (System Administrator) foresto@sun16.NoSubdomain.NoDomain (Vincent Coude' du Foresto) writes: >I ran a small testbench between a SPARC20, a PowerMac and a Quadra. The test >included computing a Fast Fourier Transform on a 1024x1024 float array, and >1000 times computing a 1024 FFT. Here are the results: >Machine FFT 1024x1024 1000 FFT 1024 >Sparc 20 8.5s 3.4s >PowerMac 7100/66 14.6s 7.4s >Centris (68040@25MHz + FPU) 39.8s 33.4s Add this data to the table 90MHz Pentium (32 MBytes RAM) 10.3s 3.2s > the built-in testbench (run in 39s by the Sparc20) On the Pentium, this runs in 17.3s. -- =============================================================== Dr. Jack Kennealy, Nashua, NH kennealy@mv.mv.com =============================================================== -------------------------------------------------- From: Chris Jacobsen Subject: Re: Platform recommendation/tradeoffs? Message-ID: <3mllmh$97r@adam.cc.sunysb.edu> Date: 14 Apr 1995 11:18:41 GMT X-Mailer: Mozilla 1.1b2 (X11; I; AIX 2) X-URL: news:ebstein-1304950926020001@lexitek.tiac.net The CPU in a pentium is within hailing distance of the CPU of new unix workstations. If that's the only issue, a PC is a good platform for IDL in terms of bang for the buck. For me, a big issue is memory. For image processing, we will frequently end up with several 1024x1024 complex floating point arrays at 8 MB memory each. We can afford to have one $10,000 unix workstation with 128 or even 256 MB of memory and a half-dozen X terminals so that one or two people can do image processing at any moment. We cannot afford to put 128 MB of memory in each of half a dozen PCs. Also, nobody ever screws up AUTOEXEC.BAT and CONFIG.SYS on the unix workstation... -------------------------------------------------- From: gurman@uvsp.gsfc.nasa.gov (Joseph B. Gurman) Subject: Re: Platform recommendation/tradeoffs? Message-ID: Date: Mon, 17 Apr 1995 18:05:32 -0400 In article <3mllmh$97r@adam.cc.sunysb.edu>, Chris Jacobsen wrote: > The CPU in a pentium is within hailing distance of the CPU of new > unix workstations. If that's the only issue, a PC is a good platform > for IDL in terms of bang for the buck. I know this isn't alt.intel.vs.the.world, but which CPU of which new workstation is which P5 within hailing distance of? A low-end SPARCstation? An AlphaStation 250 4/266? A 60 MHz P5? 100 MHz? > For me, a big issue is memory. For image processing, > we will frequently end up with several 1024x1024 complex floating > point arrays at 8 MB memory each. We can afford to have one $10,000 > unix workstation with 128 or even 256 MB of memory and a half-dozen > X terminals so that one or two people can do image processing at any > moment. We cannot afford to put 128 MB of memory in each of half a > dozen PCs. Also, nobody ever screws up AUTOEXEC.BAT and CONFIG.SYS > on the unix workstation... Actually, for a single user, the PC or Mac platform often makes more sense, simply because the memory is cheaper and the single user can do the image processing whenever he/she wants to. It's really a question of what octane, memory, and display options your application(s) require)s(. Joe Gurman -- J.B. Gurman / Solar Physics Branch/ NASA Goddard Space Flight Center/ Greenbelt MD 20771 USA / gurman@uvsp.gsfc.nasa.gov | Federal employees are still prohibited from holding opinions while| | at work. Therefore, any opinions expressed herein are somebody | | else's. | -------------------------------------------------- From: Fergus Gallagher Subject: Re: Platform recommendation/tradeoffs? Message-ID: <3n051o$nop@kwuz.nerc-keyworth.ac.uk> Date: 18 Apr 1995 10:42:00 GMT X-Mailer: Mozilla 1.1b3 (Windows; I; 16bit) I have recently taken delivery of a new SPARCstation 20, primarily for running IDL applications. I was a bit dismayed to see that my TIME_TEST benchmark was about 17.5 secs. Earlier entries in this thread have shown pentiums running better than this (15 secs, if I recall) Have I wasted my money? Or is TIME_TEST not a good measure. Obviously, my SPARCstation has more memory (64Mb + 64Mb virtual) but I could have bought this amount on memory for a pentium and still saved enough for a jolly to Bali.....in fact, I need a PC too, for running WP apps amd the like (no WABI comments, please), so I estimate I could have saved about =A38k ($12k) As I am "sole user" on my system, configuration issues aren't really a problem. Comments, anyone? Fergus | Fergus Gallagher | | Remote Sensing Applications Development Unit | | British National Space Centre | | Monks Wood | | Huntingdon PE17 2LS / UK | | | | F.Gallagher@nerc.ac.uk | | http://uh.nmt.ac.uk/bnsc/fgg.html | -------------------------------------------------- From: mallozzi@ssl.msfc.nasa.gov Subject: RE: Platform recommendation/tradeoffs? Message-ID: <3n0n1l$3fs@hammer.msfc.nasa.gov> Date: 18 Apr 1995 15:49:09 GMT >> I know this isn't alt.intel.vs.the.world, but which CPU of which new >> workstation is which P5 within hailing distance of? A low-end >> SPARCstation? An AlphaStation 250 4/266? A 60 MHz P5? 100 MHz? I ran IDL's time_test on a Pentium 90, and it took ~17 seconds. I then ran it on a DEC Alpha, and it took ~35 seconds. Grant it, the Alpha had about 40 processes going at the time, but the point is, the Pentium 90 is mighty quick. -------------------------------------------------- From: uphlabh@gemini.oscs.montana.edu (Brian Handy) Subject: Some benchmarks Message-ID: <3n3p7r$jur@pdq.coe.montana.edu> Date: 19 Apr 1995 19:44:59 GMT Keywords: Fast, fast, and...sloooow I noted the hoopla about benchmarks running on pentiums, SGI's, suns, and about every other machine, and I thought I would include some local machines for this list. This is running the TIME_TEST benchmark. SGI 4D/340S (4 CPU) 37.2 seconds {usual loaded-down state} 4D/310S (1 CPU) 39.9 {loaded} SGI Indy 17.7 {nobody} DECstation 5000/200 79.4 {nobody on} DEC Alpha (64 MB RAM) 13.3 {two other jobs} DEC Alpha (254 MB) 7.0 {nobody} The Alpha times seem to indicate this who procedure is heavily swap-dependent, which is not too big a surprise I suppose. The big SGI's aren't too revealing, as they are pretty loaded down and I didn't ever get a chance to try them otherwise. Regards, Brian -------------------------------------------------- From: velt@rad.usf.edu (Robert Velthuizen (DMIP)) Subject: Re: Platform recommendation/tradeoffs? Message-ID: <3n3qcm$lud@mother.usf.edu> Date: 19 Apr 1995 20:04:38 GMT In article , kennealy@mv.mv.com (Jack Kennealy) writes: >daft@debussy.crd.ge.com (Chris Daft) writes: > >>I second the request for some comparison of IDL's performance on >>various CPUs. Our test results: ========================== IDL ============================== IDL performance test (TIME_TEST and GRAPHICS_TIMES) Notes: (1) cd to a local disk before running the test, e.g. /tmp (2) use the second run (swap already allocated) (3) maybe do a .run time_test before running. computations graphics SS20/61/SX, 96MB, Solaris 2.3 14.2s 3.4s Solaris 2.4 13.2s SS10/51/SX, 96MB, Solaris 2.4 16.7s 4.9s SS10/dual 90MHz Hypersparc, 96MB, Solaris 2.3 13.8s 2.7s 160MB, Solaris 2.4 13.7s 2.9s SS10/51/GX, 96MB, SunOS 4.1.3 14.8s 4.0s 690MP/4proc/TGX, 128MB, SunOs 4.1.3 44.6s 7.7s SS2/1proc/GS 64MB SunOs 4.1.3 48.2s 15.2s Sun IPC/CG3 32MB SunOs 4.1.3 72.5s Robert Velthuizen (velt@rad.usf.edu), Digital Medical Imaging Program of the H. Lee Moffitt Cancer Center and Research Institute at the University of South Florida, Tampa FL 33612. -------------------------------------------------- From: gurman@uvsp.gsfc.nasa.gov (Joseph B. Gurman) Subject: Re: Some benchmarks Message-ID: Date: Wed, 19 Apr 1995 20:39:24 -0400 In article <3n3p7r$jur@pdq.coe.montana.edu>, uphlabh@gemini.oscs.montana.edu (Brian Handy) wrote: > I noted the hoopla about benchmarks running on pentiums, SGI's, suns, > and about every other machine, and I thought I would include some local > machines for this list. This is running the TIME_TEST benchmark. > > SGI 4D/340S (4 CPU) 37.2 seconds {usual loaded-down state} > 4D/310S (1 CPU) 39.9 {loaded} > SGI Indy 17.7 {nobody} > DECstation 5000/200 79.4 {nobody on} > DEC Alpha (64 MB RAM) 13.3 {two other jobs} > DEC Alpha (254 MB) 7.0 {nobody} > > The Alpha times seem to indicate this who procedure is heavily > swap-dependent, which is not too big a surprise I suppose. The big > SGI's aren't too revealing, as they are pretty loaded down and I didn't > ever get a chance to try them otherwise. Brian - There can be a lot of variables just under the designation "Alpha." For example, we see: DEC 3000 Model 300X 64 Mbyte OSF/1 3.0 15.7 s no load DEC 3000 Model 300X 64 Mbyte OSF/1 3.2 14.3 s no load DEC 3000 Model 400 64 Mbyte OVMS 6.1 21.2 s no laod DEC 3000 Model 600 288 Mbyte OSF/1 3.0 9.5 s some server jobs DEC 3000 Model 900 320 Mbyte OVMS 6.1 11.7 s no load DEC 4000 Model 710 128 Mbyte OVMS 6.1 18.8 s some server jobs all with IDL 3.6.1c. For what it's worth, there's a price range of > 3 there. For what it's worth, with an unreleased beta version of IDL for the Power Macintosh, we get: PowerMac 8100/80 48 Mbyte System 7.5.1 20.9 s Frankly, I think disk speed is a big factor in these, as the Power Mac figures (test 23 = 7.38 s) shows in comparison to say, the 3000/600 running OSF/1 (test 23 = 0.82 s). The OpenVMS systems also have extra overhead in disk writing (test 23 = 2.53 s on the 3000/900, 6.95 s on the 4000/710), but you supposedly get a more reliable file system in return. In fairness, the PowerMac disk I/O is still done in emulation. Joe P.S. Bet our PowerMac cost less than your Indy, though.... It certainly cost about a factor of 3 less than the DEC 3000/400. -- Joseph B. Gurman / NASA Goddard Space Flight Center/ Solar Data Analysis Center / Code 682 / Greenbelt MD 20771 USA / gurman@uvsp.gsfc.nasa.gov | Federal employees are still prohibited from holding opinions while at work. Any opinions expressed herein must therefore be someone else's. | -------------------------------------------------- From: sterner@strdev.jhuapl.edu (Ray Sterner) Subject: Re: Platform recommendation/tradeoffs? Message-ID: Date: Thu, 20 Apr 1995 16:20:03 GMT Sender: usenet@aplcenmp.apl.jhu.edu Nntp-Posting-Host: strdev.jhuapl.edu velt@rad.usf.edu (Robert Velthuizen (DMIP)) writes: >Our test results: >========================== IDL ============================== >IDL performance test (TIME_TEST and GRAPHICS_TIMES) >Notes: > (1) cd to a local disk before running the test, e.g. /tmp > (2) use the second run (swap already allocated) > (3) maybe do a .run time_test before running. > computations graphics >SS20/61/SX, 96MB, Solaris 2.3 14.2s 3.4s > Solaris 2.4 13.2s >SS10/51/SX, 96MB, Solaris 2.4 16.7s 4.9s >SS10/dual 90MHz Hypersparc, > 96MB, Solaris 2.3 13.8s 2.7s > 160MB, Solaris 2.4 13.7s 2.9s >SS10/51/GX, 96MB, SunOS 4.1.3 14.8s 4.0s >690MP/4proc/TGX, > 128MB, SunOs 4.1.3 44.6s 7.7s >SS2/1proc/GS 64MB SunOs 4.1.3 48.2s 15.2s >Sun IPC/CG3 32MB SunOs 4.1.3 72.5s Let me add one more machine: HP 9000/735 140Mb HP-UX A.09.05 10.1s 2.3s Ray Sterner sterner@tesla.jhuapl.edu The Johns Hopkins University North latitude 39.16 degrees. Applied Physics Laboratory West longitude 76.90 degrees. Laurel, MD 20723-6099 WWW Home page: ftp://fermi.jhuapl.edu/www/s1r/people/res/res.html -------------------------------------------------- From: kid@visdata.com (Rhonda Schienle) Subject: Re: Platform recommendation/tradeoffs? Message-ID: Date: Fri, 14 Apr 1995 19:42:17 -0800 In article <3mhc42$6lt@mozo.cc.purdue.edu>, smithjp@sage.cc.purdue.edu (Joseph Smith) wrote: > I am about to order some flavor of IDL (Mac or Unix) and > I'm wondering if anyone can help me gauge just how much performance > I loose by getting the Mac product. (This saves me enough money > to buy a Q630 cpu) Hi Joe - I've been using IDL on a Mac IIfx for the last month or so and on a Sun SPARCstation 10 for the last couple years. I also had a chance to run the native demo version of IDL on a PowerMac 7100/66 (my quick and informal testing put the PowerMac at about 75% the speed of the SS-10). The Mac IIfx runs IDL pretty well on applications that don't require a lot of memory or disk access. I have one (lots of 16-bit graphics and array manipualtions) that I was playing with today that requires about 60 MB of RAM (the IIfx only has 20 MB, so the rest was VM). As I expected, performance was pretty poor, but it ran correctly. If you already have the Q630, I think it would be adequate for what you are intending to run on it. If not, I'd get a PowerMac instead. You can download a Mac demo version (10 minute not 30 day due to Mac license managers) from gateway.rsinc.com to try out on your Q630 to get a decent idea. -- Rhonda Schienle Email: kid@visdata.com Licensed Massage Therapist on California's Central Coast -------------------------------------------------- From: ps@kis.uni-freiburg.de (Peter Suetterlin) Subject: Re: Platform recommendation/tradeoffs? Message-ID: <3mp0l2$7or@n.ruf.uni-freiburg.de> Date: 15 Apr 1995 17:44:02 GMT X-Newsreader: TIN [version 1.2 PL2] Chris Jacobsen (jacobsen@xray1.physics.sunysb.edu) wrote: : The CPU in a pentium is within hailing distance of the CPU of new : unix workstations. If that's the only issue, a PC is a good platform : for IDL in terms of bang for the buck. : For me, a big issue is memory. For image processing, : we will frequently end up with several 1024x1024 complex floating : point arrays at 8 MB memory each. We can afford to have one $10,000 : unix workstation with 128 or even 256 MB of memory and a half-dozen : X terminals so that one or two people can do image processing at any : moment. We cannot afford to put 128 MB of memory in each of half a : dozen PCs. Also, nobody ever screws up AUTOEXEC.BAT and CONFIG.SYS : on the unix workstation... So get the PC, put 128MB and Linux to it. IDL 4.0 is announced for Linux Then connect your XTerms to it and enjoy. Peter ------------------ Peter 'PIT' Suetterlin ----------------- | Kiepenheuer Institut | Sternfreunde Breisgau e.V | | fuer Sonnenphysik | | | 0761/3198-210 | 0761/71571 | ---- -------------------------------------------------- From: ps@kis.uni-freiburg.de (Peter Suetterlin) Subject: Re: Platform recommendation/tradeoffs? Message-ID: <3ng9ck$35e@n.ruf.uni-freiburg.de> Date: 24 Apr 1995 13:34:12 GMT X-Newsreader: TIN [version 1.2 PL2] Ray Sterner (sterner@strdev.jhuapl.edu) wrote: : velt@rad.usf.edu (Robert Velthuizen (DMIP)) writes: : >Our test results: : >========================== IDL ============================== : >IDL performance test (TIME_TEST and GRAPHICS_TIMES) : >Notes: : > (1) cd to a local disk before running the test, e.g. /tmp : > (2) use the second run (swap already allocated) : > (3) maybe do a .run time_test before running. : > computations graphics : >SS20/61/SX, 96MB, Solaris 2.3 14.2s 3.4s : > Solaris 2.4 13.2s : >SS10/51/SX, 96MB, Solaris 2.4 16.7s 4.9s : >SS10/dual 90MHz Hypersparc, : > 96MB, Solaris 2.3 13.8s 2.7s : > 160MB, Solaris 2.4 13.7s 2.9s : >SS10/51/GX, 96MB, SunOS 4.1.3 14.8s 4.0s : >690MP/4proc/TGX, : > 128MB, SunOs 4.1.3 44.6s 7.7s : >SS2/1proc/GS 64MB SunOs 4.1.3 48.2s 15.2s : >Sun IPC/CG3 32MB SunOs 4.1.3 72.5s : Let me add one more machine: : HP 9000/735 140Mb HP-UX A.09.05 10.1s 2.3s Allthough maybe not completely comparable, as it is PV-WAVE (still waiting for the IDL-Version) using the time_test of IDL 3.61a: 486DX/2-80 16MB Linux 1.2.5 30.2s 7.8s (1MB S3 card 8bit) ------------------ Peter 'PIT' Suetterlin ----------------- | Kiepenheuer Institut | Sternfreunde Breisgau e.V | | fuer Sonnenphysik | | | 0761/3198-210 | 0761/71571 | ---- -------------------------------------------------- From: clodius@lanl.gov (William B. Clodius) Subject: Re: Some benchmarks Message-ID: Date: Tue, 25 Apr 95 21:22:58 GMT X-Newsreader: VersaTerm Link v1.1.1 In Article , gurman@uvsp.gsfc.nasa.gov (Joseph B. Gurman) wrote: > > For what it's worth, with an unreleased beta version of IDL for the Power >Macintosh, we get: > > PowerMac 8100/80 48 Mbyte System 7.5.1 20.9 s > >Frankly, I think disk speed is a big factor in these, as the Power Mac figures >(test 23 = 7.38 s) shows in comparison to say, the 3000/600 running OSF/1 >(test 23 = 0.82 s). The OpenVMS systems also have extra overhead in disk >writing (test 23 = 2.53 s on the 3000/900, 6.95 s on the 4000/710), but >you supposedly get a more reliable file system in return. In fairness, the >PowerMac disk I/O is still done in emulation. > > Joe > >P.S. Bet our PowerMac cost less than your Indy, though.... It certainly >cost about a factor of 3 less than the DEC 3000/400. > >-- >Joseph B. Gurman / NASA Goddard Space Flight Center/ Solar Data Analysis Center / Code 682 / Greenbelt MD 20771 USA / gurman@uvsp.gsfc.nasa.gov >| Federal employees are still prohibited from holding opinions while at work. Any opinions expressed herein must therefore be someone else's. | Be aware that time_test for the Power Macs, and probably for a lot of other processors is dependent on more than just processor, raw memory, clock speed, and disk speed. I had an opportunity to try out IDL in a class on a Power Mac 6100 with 16 Mbytes of memory. Experimenting with time_test showed that it also depended on whether 1. It was run the first or latter time. On the first time you had additional disk accesses and compilation to deal with so that the second run was significantly faster. 2. The disk cache and program had adequate memory set aside (critical on a 16 Mbyte machine as the version of IDL reccomends 10 Mbytes + for IDL alone). If the cache was large enough, disk access was less of a problem. 3. Graphics were written to a buffer for updating. Without the buffer you had some problems with the appearance of windows ;-), but about twice the nominal graphics performance. 4. Modern Memory manager and Virtual memory were used. 5. Sometimes performance would change with configuration in a nonobvious way, resulting in a factor of four or more degradation in performance. 6. Whether the beta or older non beta version was used. The beta version was typically more stable and faster than the nonbeta version, but could not run a few of the files. For the record, with graphics buffering, a beta version of IDL on the 6100 Power PC, for the first run of time test I remember getting results between 30 and 70 s, for the second run between 17 and 70 s. The non-beta version was 25-50% slower. A 90 MHz Pentium nearby was reporting about 16 s, and over twice the graphics performance. Most of the performance was determined by the disk intensive routines, which varied by over an order of magnitude depending on the cache, etc. settings. I recall test 23 taking 3-4 seconds with the optimum setting on the second run, and more than 20 seconds on the worst setting. In the class itself, where routines were applied to arrays that could never fit in the disk cache on a 16 Mbyte machine, performance was at best a factor of two worse than the Suns and PCs. It was also significantly less stable than the Suns with most of the problems due to memory conditions, but also had problems with directories on CD-ROMs where it would try to perform a write operation ;-) andd bomb instead of failing gracefully. It appeared to also be less stable than the PCs, but the PCs also bombed a number of times. I do not recomend running IDL on a 16 Mbyte Power Mac, 24 Mbytes may be adequate, but you can never have too much memory. ------------------------------------------------------------------------ William B. Clodius, wclodius@lanl.gov NIS-1, Los Alamos National Laboratory ------------------------------------------------------------------------