Jul/2020 Update! I have redone all Indy, Indigo2, O2, Octane and most IRIS Indigo tests using exactly the same model disk (Seagate 36GB 15K ST336753LC) and where possible the same compiler version, so the data is now much more comparable than the old results. I have obtained a SCSI2SD V6 card and will try that out later in targeted systems, along with one or two SATA SSDs using a SCSI/SATA bridge box. In the future I might also add just a few specific O2/Octane tests using the Fujitsu MAS3367NC and MAX3036NC respectively, since these are the disk models I normally use and supply in such systems.
Results submitted by 3rd parties may vary, so compare with caution.
There are three tables, for the following tests:
Compiler References: {1} MIPS Pro 7.4.4m {2} MIPS Pro 7.3 (7.4.1m) {3} GCC 3.4.6 (Nekoware build) {4} MIPS Pro 7.3 (7.3) {5} From IDO 5.3 for IRIX 5.3 {6} GCC 2.6.3 (SGI Freeware for IRIX 5.3) {7} MIPS Pro 7.3 (7.4.3m) Disk References: {1} Seagate 36GB 15K ST336753LC (BF03685A35) {2} Fujitsu MAS3367NC 36GB 15K (Worldisk MX03367SC800600S) {3} Fujitsu MAX3036NC 36GB 15K (Worldisk MX03036SC800600X) {4} Seagate Cheetah ST336704LCV U160 36GB 1" 10000rpm SCA {5} IBM-ESXSBBD073C3ESTT0ZFNJP86 73GB 15000rpm SCA (Maxtor Atlas 15K II) {6} COMPAQ BD1468A4B5 146GB 10K SCA {7} Fujitsu MAX3036NP 36GB 15000rpm 68pin {8} Seagate 18GB 15000rpm SCA (COMPAQ BD01862376) {9} Seagate ST3300655LC 300GB 15000rpm SCA MIPS Pro (-O): CPU No. L2/L3 IRIX Compiler Disk Time System Type CPUs MHz Cache Version Ref Type mm:ss NOTES Tezro R16000 4 1000 16MB 6.5.26m 7 4 00:42 Fuel R16000 1 900 8MB 6.5.26m 7 4 00:43 Fuel R16000 1 800 4MB 6.5.26m 7 4 00:48 Tezro R16000 4 700 4MB 6.5.26m 7 4 00:53 Fuel R16000 1 700 4MB 6.5.26m 7 4 00:54 Origin350 R16000 4 700 4MB 6.5.26m 4 6 00:56 Fuel R14000 1 600 4MB 6.5.26m 7 4 01:04 Origin300 R14000 2 600 4MB 6.5.30m 1 1 01:14 Origin300 R14000 2 500 2MB 6.5.26m 7 8 01:16 Octane R14000 1 550 2MB 6.5.26m 7 5 01:17 Octane R14000 2 600 2MB 6.5.30m 1 1 01:22 Octane R12000 2 400 2MB 6.5.30m 1 1 01:53 Just to prove only one CPU is used. A tiny gain from general OS speedup. Octane R12000 1 400 2MB 6.5.30m 1 1 01:54 Octane R12000 1 360 2MB 6.5.30m 1 1 02:11 Octane R12000 1 300 2MB 6.5.30m 1 1 02:28 Octane R12000 1 270 2MB 6.5.30m 1 1 02:44 Octane R10000 1 250 2MB 6.5.30m 1 1 02:56 CPU mod; was an R12K/300 2MB, core swapped with an R10K/250 for O2 and clock changed to 250. O2 R7000 1 600 256K/1MB 6.5.30m 1 1 02:58 O2 R12000 1 400 2MB 6.5.30m 1 1 03:05 Octane R10000 1 250 1MB 6.5.30m 1 1 03:05 Onyx R10000 4 195 2MB 6.5.22m 2 4 03:13 Octane R10000 1 225 1MB 6.5.30m 1 1 03:46 Octane R10000 1 195 1MB 6.5.30m 1 1 03:47 O2 R7000 1 350 256K/1MB 6.5.30m 1 1 04:10 O2 R12000 1 300 1MB 6.5.30m 1 1 04:23 Indigo2 R10000 1 195 1MB 6.5.22m 1 1 04:38 Octane R10000 1 175 1MB 6.5.30m 1 1 04:43 O2 R12000 1 270 2MB 6.5.30m 1 1 04:49 O2 R10000 1 250 1MB 6.5.30m 1 1 04:52 Indigo2 R10000 1 175 1MB 6.5.22m 1 1 04:59 O2 R10000 1 195 1MB 6.5.30m 1 1 05:43 Onyx R4400SC 1 250 4MB 6.5.22m 2 4 05:48 O2 R5200 1 300 1MB 6.5.30m 1 1 05:51 O2 R10000 1 175 1MB 6.5.30m 1 1 07:02 O2 R5000SC 1 200 1MB 6.5.30m 1 1 07:39 Indigo2 R4400SC 1 250 2MB 6.5.22m 1 1 07:39 O2 R10000 1 150 1MB 6.5.30m 1 1 07:56 Onyx R4400SC 4 200 4MB 6.5.22m 1 9 07:59 Disk connected via SE bus (10MB/sec) O2 R5000SC 1 180 512K 6.5.30m 1 1 08:58 Indigo2 R4400SC 1 200 2MB 6.5.22m 1 1 09:48 Indy R5000SC 1 180 512K 6.5.22m 1 1 09:57 Indy R5000SC 1 150 512K 6.5.22m 1 1 10:29 Indigo2 R4400SC 1 200 1MB 6.5.22m 1 1 10:55 Indy R4400SC 1 200 1MB 6.5.22m 1 1 11:03 IRIS Indigo R4400SC 1 150 1MB 6.5.22m 1 1 13:39 Indy R4400SC 1 150 1MB 6.5.22m 1 1 13:52 Indy R4600SC 1 133 512K 6.5.22m 1 1 14:14 Indigo2 R4400SC 1 150 1MB 6.5.22m 1 1 14:18 Indy R5000PC 1 150 - 6.5.22m 1 1 14:44 IRIS Indigo R3000 1 33 - 5.3 5 1 15:27 Fascinating. :D O2 R5000PC 1 180 - 6.5.30m 1 1 15:37 IRIS Indigo R4000SC 1 100 1MB 6.5.22m 1 1 21:41 Indy R4000SC 1 100 1MB 6.5.22m 1 1 21:48 Indy R4600PC 1 133 - 6.5.22m 1 1 22:55 Indy R4600PC 1 100 - 6.5.22m 1 1 23:36 Personal IRIS R2000 1 12 - 5.3 5 7 51:35 IP6 motherboard. ******************************************************************************************************************************************************************** Disk Time Compiler MIPS Pro (-O0): Ref mm:ss Reference CPU No. L2/L3 IRIX Compiler Disk Time System Type CPUs MHz Cache Version Ref Type mm:ss NOTES Tezro R16000 4 1000 16MB 6.5.26m 7 4 00:18 Fuel R16000 1 900 8MB 6.5.26m 7 4 00:19 Fuel R16000 1 800 4MB 6.5.26m 7 4 00:21 Fuel R16000 1 700 4MB 6.5.26m 7 4 00:23 Tezro R16000 4 700 4MB 6.5.26m 7 4 00:24 Origin350 R16000 4 700 4MB 6.5.30m 4 6 00:25 Fuel R14000 1 600 4MB 6.5.26m 7 4 00:27 Origin300 R14000 2 600 4MB 6.5.30m 1 1 00:33 Origin300 R14000 2 500 2MB 6.5.26m 7 8 00:33 Octane R14000 1 550 2MB 6.5.26m 7 5 00:33 Octane R14000 2 600 2MB 6.5.30m 1 1 00:35 Octane R12000 2 400 2MB 6.5.30m 1 1 00:49 Just to prove only one CPU is used. A tiny gain from general OS speedup. Octane R12000 1 400 2MB 6.5.30m 1 1 00:50 Octane R12000 1 360 2MB 6.5.30m 1 1 00:58 Octane R12000 1 300 2MB 6.5.30m 1 1 01:05 O2 R7000 1 600 256K/1MB 6.5.30m 1 1 01:11 Octane R12000 1 270 2MB 6.5.30m 1 1 01:13 Octane R10000 1 250 2MB 6.5.30m 1 1 01:18 CPU mod; was an R12K/300 2MB, core swapped with an R10K/250 for O2 and clock changed to 250. O2 R12000 1 400 2MB 6.5.30m 1 1 01:18 Octane R10000 1 250 1MB 6.5.30m 1 1 01:22 Onyx R10000 4 195 2MB 6.5.22m 2 4 01:24 Octane R10000 1 225 1MB 6.5.30m 1 1 01:42 O2 R7000 1 350 256K/1MB 6.5.30m 1 1 01:43 Octane R10000 1 195 1MB 6.5.30m 1 1 01:52 O2 R12000 1 300 1MB 6.5.30m 1 1 01:53 Indigo2 R10000 1 195 1MB 6.5.22m 1 1 02:02 O2 R12000 1 270 2MB 6.5.30m 1 1 02:05 Octane R10000 1 175 1MB 6.5.30m 1 1 02:06 O2 R10000 1 250 1MB 6.5.30m 1 1 02:07 Indigo2 R10000 1 175 1MB 6.5.22m 1 1 02:13 O2 R10000 1 195 1MB 6.5.30m 1 1 02:30 O2 R5200 1 300 1MB 6.5.30m 1 1 02:33 Onyx R4400SC 1 250 4MB 6.5.22m 2 4 02:41 O2 R10000 1 175 1MB 6.5.30m 1 1 03:02 O2 R5000SC 1 200 1MB 6.5.30m 1 1 03:22 O2 R10000 1 150 1MB 6.5.30m 1 1 03:28 Onyx R4400SC 4 200 4MB 6.5.22m 1 9 03:35 Disk connected via SE bus (10MB/sec) Indigo2 R4400SC 1 250 2MB 6.5.22m 1 1 03:55 O2 R5000SC 1 180 512K 6.5.30m 1 1 03:56 Indigo2 R4400SC 1 200 2MB 6.5.22m 1 1 04:26 Indy R5000SC 1 180 512K 6.5.22m 1 1 04:27 Indy R5000SC 1 150 512K 6.5.22m 1 1 04:38 Indy R4400SC 1 200 1MB 6.5.22m 1 1 04:54 Indigo2 R4400SC 1 200 1MB 6.5.22m 1 1 05:15 Indy R4400SC 1 150 1MB 6.5.22m 1 1 05:58 IRIS Indigo R4400SC 1 150 1MB 6.5.22m 1 1 06:08 Indy R4600SC 1 133 512K 6.5.22m 1 1 06:28 Indigo2 R4400SC 1 150 1MB 6.5.22m 1 1 06:35 Indy R5000PC 1 150 - 6.5.22m 1 1 06:48 O2 R5000PC 1 180 - 6.5.30m 1 1 07:14 IRIS Indigo R3000 1 33 - 5.3 5 1 08:50 Fascinating. :D Indy R4000SC 1 100 1MB 6.5.22m 1 1 09:54 IRIS Indigo R4000SC 1 100 1MB 6.5.22m 1 1 09:57 Indy R4600PC 1 133 - 6.5.22m 1 1 10:58 Indy R4600PC 1 100 - 6.5.22m 1 1 11:08 Personal IRIS R2000 1 12 - 5.3 5 7 31:57 IP6 motherboard. ******************************************************************************************************************************************************************** GCC: Disk Time Compiler Ref mm:ss Reference CPU No. L2/L3 IRIX Compiler Disk Time System Type CPUs MHz Cache Version Ref Type h:mm:ss NOTES Tezro R16000 4 1000 16MB 6.5.26m 3 4 0:00:43 Fuel R16000 1 900 8MB 6.5.26m 3 4 0:00:44 Fuel R16000 1 800 4MB 6.5.26m 3 4 0:00:49 Tezro R16000 4 700 4MB 6.5.26m 3 4 0:00:54 Fuel R16000 1 700 4MB 6.5.26m 3 4 0:00:54 Fuel R14000 1 600 4MB 6.5.26m 3 4 0:01:04 Origin300 R14000 2 600 4MB 6.5.30m 3 1 0:01:04 Octane R14000 2 600 2MB 6.5.30m 3 1 0:01:07 Origin300 R14000 2 500 2MB 6.5.26m 3 8 0:01:15 Octane R14000 1 550 2MB 6.5.26m 3 5 0:01:15 Octane R12000 2 400 2MB 6.5.30m 3 1 0:01:34 Just to prove only one CPU is used. A tiny gain from general OS speedup. Octane R12000 1 400 2MB 6.5.30m 3 1 0:01:35 Octane R12000 1 360 2MB 6.5.30m 3 1 0:01:51 Octane R12000 2 350 1MB 6.5.26m 3 4 0:01:55 CPU mod. O2 R7000 1 600 256K/1MB 6.5.30m 3 1 0:02:04 Octane R12000 1 300 2MB 6.5.30m 3 1 0:02:05 O2 R12000 1 400 2MB 6.5.30m 3 1 0:02:18 Octane R12000 1 270 2MB 6.5.30m 3 1 0:02:19 Octane R10000 1 250 2MB 6.5.30m 3 1 0:02:29 CPU mod; was an R12K/300 2MB, core swapped with an R10K/250 for O2 and clock changed to 250. Octane R10000 1 250 1MB 6.5.30m 3 1 0:02:35 O2 R7000 1 350 256K/1MB 6.5.30m 3 1 0:03:03 Octane R10000 1 225 1MB 6.5.30m 3 1 0:03:09 O2 R12000 1 300 1MB 6.5.30m 3 1 0:03:18 Octane R10000 1 195 1MB 6.5.30m 3 1 0:03:29 O2 R12000 1 270 2MB 6.5.30m 3 1 0:03:38 Indigo2 R10000 1 195 1MB 6.5.22m 3 1 0:03:39 O2 R10000 1 250 1MB 6.5.30m 3 1 0:03:46 Octane R10000 1 175 1MB 6.5.30m 3 1 0:03:55 Indigo2 R10000 1 175 1MB 6.5.22m 3 1 0:03:58 O2 R5200 1 300 1MB 6.5.30m 3 1 0:04:18 O2 R10000 1 195 1MB 6.5.30m 3 1 0:04:23 O2 R10000 1 175 1MB 6.5.30m 3 1 0:05:19 O2 R5000SC 1 200 1MB 6.5.30m 3 1 0:05:48 O2 R10000 1 150 1MB 6.5.30m 3 1 0:06:02 Onyx R4400SC 4 200 4MB 6.5.22m 3 9 0:06:25 Disk connected via SE bus (10MB/sec) Indigo2 R4400SC 1 250 2MB 6.5.22m 3 1 0:06:41 O2 R5000SC 1 180 512K 6.5.30m 3 1 0:06:47 Indy R5000SC 1 180 512K 6.5.22m 3 1 0:07:20 Indigo2 R4400SC 1 200 2MB 6.5.22m 3 1 0:07:39 Indy R5000SC 1 150 512K 6.5.22m 3 1 0:07:54 Indy R4400SC 1 200 1MB 6.5.22m 3 1 0:08:21 Indigo2 R4400SC 1 200 1MB 6.5.22m 3 1 0:08:31 IRIS Indigo R4400SC 1 150 1MB 6.5.22m 3 1 0:10:41 Indy R4400SC 1 150 1MB 6.5.22m 3 1 0:10:43 Indy R5000PC 1 150 - 6.5.22m 3 1 0:10:46 Indy R4600SC 1 133 512K 6.5.22m 3 1 0:10:49 Indigo2 R4400SC 1 150 1MB 6.5.22m 3 1 0:11:09 O2 R5000PC 1 180 - 6.5.30m 3 1 0:11:18 IRIS Indigo R3000 1 33 - 5.3 5 1 0:16:10 Fascinating. :D Indy R4600PC 1 133 - 6.5.22m 3 1 0:17:04 IRIS Indigo R4000SC 1 100 1MB 6.5.22m 3 1 0:17:17 Indy R4000SC 1 100 1MB 6.5.22m 3 1 0:17:49 Indy R4600PC 1 100 - 6.5.22m 3 1 0:17:59 Personal IRIS R2000 1 12 - 5.3 6 7 1:01:47 IP6 motherboard.
If you want to perform this test yourself, download the archive and then follow these instructions:
As root, decompress and untar the archive into /var/tmp, then cd to the lynxtest directory and execute the RUN scripts one at a time, ie.:
gunzip lynxtest.tar.gz tar xvf lynxtest.tar cd lynxtest ./RUN1 ./RUN2 ./RUN2
Each script will display the required time to submit at the end of the test. If you send me results for your own system, please include a full description of the system, ie. 'hinv -vm' output, and the results of the following three commands so that I can tell which versions of the compilers were used and what model disk the tests are being run with:
cc -version gcc --version scsicontrol -i /dev/scsi/sc0d1l0
The above assumes the system disk is on controller 0, ID 1. If this is not the case (eg. R10K/R12K O2 use SCSI ID 2) then adjust the scsicontrol command accordingly.
In early testing I found that the old 7.3 MIPS Pro compiler is much faster, enough to make an R12K/400 Octane using 7.4.4 complete a compilation about as fast as an R12K/300 Octane does with 7.3. Of course it may be that 7.4.4 is producing better binaries, in which case final code execution speed may be better, but this test isn't about that. In time I may include a separate test using one specific system to show how compilation speed varies based on compiler version.
Grud knows why the R3K Indigo is so comparatively fast.
Lastly, note that O2 systems may be slightly faster when operating with lower display resolutions and/or bit depths.