| Title: | FDDI - The Next Generation |
| Moderator: | NETCAD::STEFANI |
| Created: | Thu Apr 27 1989 |
| Last Modified: | Thu Jun 05 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 2259 |
| Total number of notes: | 8590 |
Where can I find a figure for the sustainable throughput of a FDDI on a VAXstation 40000 Model 90? Wes
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 754.1 | STAR::GAGNE | David Gagne - VMS/LAN Development | Tue Oct 20 1992 08:35 | 4 | |
Are you interested in what a particular protocol can do or are you
interested in what the VMS driver can do by itself?
Also, I assume you want the throughput of the DEFZA on the 4000-90.
| |||||
| 754.2 | MIPSBX::thomas | The Code Warrior | Tue Oct 20 1992 10:50 | 1 | |
I think the DEFTA's numbers would be much more interesting. | |||||
| 754.3 | NLA0::ONO | The Wrong Stuff | Tue Oct 20 1992 12:16 | 5 | |
DECnet FAL, DECnet intertask, bare VMS driver, if you have them all. The customer is looking for about 5 MB/s to support a simulation. Wes | |||||
| 754.4 | STAR::GAGNE | David Gagne - VMS/LAN Development | Tue Oct 20 1992 15:37 | 4 | |
re: .2 - We are developing the DEFTA driver now, so we don't have
numbers for it yet.
I am looking for the numbers for the DEFZA now.
| |||||
| 754.5 | Here's some data for you | STAR::GAGNE | David Gagne - VMS/LAN Development | Tue Oct 20 1992 17:54 | 121 |
SELF/1 XX> B0 XX> T
SELF/1 XX> B0 X
Transmt test results at 7-AUG-1992 10:31:53.96 (Completed):
Test node .......... COUGR5 on 1/FCA0 (FDDI)
Test duration ...... 7-AUG-1992 10:30:32.80 => 7-AUG-1992 10:31:53.89 for 0:01:21.090
Test parameters .... ETH, Pipeline 4, Chaining Off, User mask 00000001, Duration 10, Target node NMSMON
CPU type ........... VAXstation 4000-90, 1 CPU active
Versions ........... Sys VMS-T5.5-2X4L, Drv 05052007 00000001, Lan 05052025 A4000001, Hw 00000000
System parameters .. Multiprocessing 0, Poolcheck 0, Vaxcluster 0
-----22--- -----65--- -----96--- ----128--- ----192--- ----512--- ---1518--- ---4495---
Test time .......... 10.000 10.000 10.000 10.000 10.000 10.000 10.000 10.000
Intr time .......... 100% 100% 100% 100% 100% 100% 100% 100%
Xmt (success) ...... 214414 143707 114133 94530 70624 31166 11085 3820
per second ...... 21441 14371 11413 9453 7062 3117 1109 382
Xmt total bytes .... 5360350 9772076 11299167 12383430 13771680 16050490 16860286 17182360
per second ...... 536035 977208 1129917 1238343 1377168 1605049 1686029 1718236
Xmt user bytes ..... 4294324054 5748280 8103443 9736590 11794208 15177842 16549905 17075400
per second ...... 4294902972 574828 810344 973659 1179421 1517784 1654990 1707540
SELF/1 XX> B0 XX> $ SET TERM/
SELF/1 XX> B0 XX> T
SELF/1 XX> B0 X
Transmt test results at 7-AUG-1992 10:33:39.01 (Completed):
Test node .......... COUGR5 on 1/FCA0 (FDDI)
Test duration ...... 7-AUG-1992 10:32:18.85 => 7-AUG-1992 10:33:38.93 for 0:01:20.080
Test parameters .... ETH, Pipeline 4, Chaining Off, User mask 00000001, Duration 10, Target node NMSMON
CPU type ........... VAXstation 4000-90, 1 CPU active
Versions ........... Sys VMS-T5.5-2X4L, Drv 05052007 00000001, Lan 05052025 A4000001, Hw 00000000
System parameters .. Multiprocessing 0, Poolcheck 0, Vaxcluster 0
-----22--- -----65--- -----96--- ----128--- ----192--- ----512--- ---1518--- ---4495---
Test time .......... 10.000 10.000 10.000 10.000 10.000 10.000 10.000 10.000
Intr time .......... 100% 100% 100% 100% 100% 100% 100% 100%
Xmt (success) ...... 213788 143424 114001 94493 70565 31138 11081 3819
per second ...... 21379 14342 11400 9449 7057 3114 1108 382
Xmt total bytes .... 5344700 9752832 11286099 12378583 13760175 16036070 16854202 17177862
per second ...... 534470 975283 1128610 1237858 1376018 1603607 1685420 1717786
Xmt user bytes ..... 4294325932 5736960 8094071 9732779 11784355 15164206 16543933 17070930
per second ...... 4294903160 573696 809407 973278 1178436 1516421 1654393 1707093
SELF/1 XX> B0 XX> T
SELF/1 XX> B0 X
Transmt test results at 7-AUG-1992 10:36:01.08 (Completed):
Test node .......... COUGR5 on 1/FCA0 (FDDI)
Test duration ...... 7-AUG-1992 10:34:40.91 => 7-AUG-1992 10:36:01.00 for 0:01:20.090
Test parameters .... ETH, Pipeline 4, Chaining Off, User mask 00000001, Duration 10, Target node NMSMON
CPU type ........... VAXstation 4000-90, 1 CPU active
Versions ........... Sys VMS-T5.5-2X4L, Drv 05052007 00000001, Lan 05052025 A4000001, Hw 00000000
System parameters .. Multiprocessing 0, Poolcheck 0, Vaxcluster 0
-----22--- -----65--- -----96--- ----128--- ----192--- ----512--- ---1518--- ---4495---
Test time .......... 10.000 10.000 10.000 10.000 10.000 10.000 10.000 10.000
Intr time .......... 100% 100% 100% 100% 100% 100% 100% 100%
Xmt (success) ...... 214024 143551 114048 94501 70604 31160 11085 3819
per second ...... 21402 14355 11405 9450 7060 3116 1109 382
Xmt total bytes .... 5350600 9761468 11290752 12379631 13767780 16047400 16860286 17177862
per second ...... 535060 976147 1129075 1237963 1376778 1604740 1686029 1717786
Xmt user bytes ..... 4294325224 5742040 8097408 9733603 11790868 15174920 16549905 17070930
per second ...... 4294903089 574204 809741 973360 1179087 1517492 1654990 1707093
SELF/1 XX> B0 XX> SET SIZE E
%XX-E-INVAL, Invalid command, ETHERNET
SELF/1 XX> B0 XX> SET SIZE E
%XX-E-INVAL, Invalid command, ETHERNET
SELF/1 XX> B0 XX> HELP S
SELF/1 XX> B0 XX> SET SIZE
SELF/1 XX> B0 XX>
Parameters:
Log output file .... COUGAR_FDDI_XX.OUT
Cmd file echo mode . Off
Wait for completion No
Print per/sec also . Yes
Allow on public ntwk No
Test node .......... SELF/1
Test parameters .... ETH, Pipeline 4, Chaining Off, User mask 00000000, Duration 10, Target node NMSMON/1
Traffic parameters . 0 64-byte packets at 0/sec
Packet sizes ....... 64 65 96 128 192 512 1024 1518
SELF/1 XX> B0 XX> T
SELF/1 XX> B0 X
Transmt test results at 7-AUG-1992 10:42:14.25 (Completed):
Test node .......... COUGR5 on 1/FCA0 (FDDI)
Test duration ...... 7-AUG-1992 10:40:54.08 => 7-AUG-1992 10:42:14.17 for 0:01:20.090
Test parameters .... ETH, Pipeline 4, Chaining Off, User mask 00000001, Duration 10, Target node NMSMON
CPU type ........... VAXstation 4000-90, 1 CPU active
Versions ........... Sys VMS-T5.5-2X4L, Drv 05052007 00000001, Lan 05052025 A4000001, Hw 00000000
System parameters .. Multiprocessing 0, Poolcheck 0, Vaxcluster 0
-----64--- -----65--- -----96--- ----128--- ----192--- ----512--- ---1024--- ---1518---
Test time .......... 10.000 10.000 10.000 10.000 10.000 10.000 10.000 10.000
Intr time .......... 100% 100% 100% 100% 100% 100% 100% 100%
Xmt (success) ...... 143508 143678 114099 94531 70614 31144 16228 11085
per second ...... 14351 14368 11410 9453 7061 3114 1623 1109
Xmt total bytes .... 9615036 9770104 11295801 12383561 13769730 16039160 16666156 16860286
per second ...... 961504 977010 1129580 1238356 1376973 1603916 1666616 1686029
Xmt user bytes ..... 5596812 5747120 8101029 9736693 11792538 15167128 16211772 16549905
per second ...... 559681 574712 810103 973669 1179254 1516713 1621177 1654990
SELF/1 XX> B0 XX> T
SELF/1 XX> B0 X
Transmt test results at 7-AUG-1992 10:44:54.32 (Completed):
Test node .......... COUGR5 on 1/FCA0 (FDDI)
Test duration ...... 7-AUG-1992 10:43:34.15 => 7-AUG-1992 10:44:54.24 for 0:01:20.090
Test parameters .... ETH, Pipeline 4, Chaining Off, User mask 00000001, Duration 10, Target node NMSMON
CPU type ........... VAXstation 4000-90, 1 CPU active
Versions ........... Sys VMS-T5.5-2X4L, Drv 05052007 00000001, Lan 05052025 A4000001, Hw 00000000
System parameters .. Multiprocessing 0, Poolcheck 0, Vaxcluster 0
-----64--- -----65--- -----96--- ----128--- ----192--- ----512--- ---1024--- ---1518---
Test time .......... 10.000 10.000 10.000 10.000 10.000 10.000 10.000 10.000
Intr time .......... 100% 100% 100% 100% 100% 100% 100% 100%
Xmt (success) ...... 143599 143597 114072 94490 70625 31152 16233 11085
per second ...... 14360 14360 11407 9449 7063 3115 1623 1109
Xmt total bytes .... 9621133 9764596 11293128 12378190 13771875 16043280 16671291 16860286
per second ...... 962113 976460 1129313 1237819 1377188 1604328 1667129 1686029
Xmt user bytes ..... 5600361 5743880 8099112 9732470 11794375 15171024 16216767 16549905
per second ...... 560036 574388 809911 973247 1179438 1517103 1621677 1654990
SELF/1 XX> B0 X
| |||||
| 754.6 | STAR::GAGNE | David Gagne - VMS/LAN Development | Tue Oct 20 1992 17:57 | 8 | |
The information in .4 is the raw throughput for the DEFZA. I have no
numbers for DECnet.
Also, the numbers may be better than what you see - the driver was not
yet done (optimized) when those numbers were taken. The numbers will
vary will the addition of a user, with some other piece of code using
the VAX CPU, with the version of VMS, with the CPU it runs on, and with
the version of the hardware. (Your mileage may vary.)
| |||||
| 754.7 | SUBWAY::BRIGGS | Have datascope, will travel. | Wed Oct 21 1992 14:17 | 2 | |
I am confused. What does 4,294,325,932 bytes/second under the 22 column
indicate?
| |||||
| 754.8 | bytes/second for a stream of 22 byte packets | DECEAT::KALKUNTE | Ramsesh Kalkunte 293-5139 | Wed Oct 21 1992 16:38 | 1 |
| 754.9 | still confused | ERICG::ERICG | Eric Goldstein | Thu Oct 22 1992 03:30 | 9 |
.7> I am confused. What does 4,294,325,932 bytes/second under the 22 column .7> indicate? .8> -< bytes/second for a stream of 22 byte packets >- It's very impressive that you're moving the data from somewhere to somewhere else at more than 4 gigabytes per second, but I find it hard to believe that it's actually going down the wire that fast. Could we please have some further clarification? Just what are "Xmt user bytes per second"? | |||||
| 754.10 | it means very little | MUDDY::WATERS | Thu Oct 22 1992 05:35 | 4 | |
The 4 GB/s figure is caused by underrun in the program's arithmetic.
A packet's "user bytes" is defined as "total bytes" minus 28, to
account for protocol overhead. 22 minus 28 equals 4 billion in C,
as it happens. 8^(.
| |||||
| 754.11 | I should have included a :-) in my earlier note | DECEAT::KALKUNTE | Ramsesh Kalkunte 293-5139 | Thu Oct 22 1992 12:06 | 9 |
yes, this is due to arithmetic error in the program and no, it is not
specific to C. If you declare 32 bit arithmetic functions as "unsigned"
and you have negative values generated from the function, this is what
you get. Why it happens is something I will leave as an exercise to the
readerr. So, disregard the numbers for 22 byte packets. For other
packet sizes, this gives you a good feel for how much of network user
data you can push through the wire.
Ram
| |||||
| 754.12 | well, er, that's really exciting.... | NLA0::ONO | The Wrong Stuff | Thu Oct 22 1992 17:30 | 3 |
1.6 MB/s, eh? Gee, that's not as fast as I expected. Wes | |||||
| 754.13 | STAR::GAGNE | David Gagne - VMS/LAN Development | Fri Oct 23 1992 12:17 | 4 | |
On AXP, the numbers are much better because the driver has been
optimized a little more and because we have more CPU cycles. I believe
that you run out of CPU cycles on the VAX that the numbers were
generated on.
| |||||
| 754.14 | Just what IS a DEFTA?? | DYOSW8::WILDER | Does virtual reality get swapped? | Tue Jan 12 1993 14:09 | 7 |
Okay, I have to ask....I have looked in this conference and cannot find
the answer.....
What is the DEFTA...is this the Q-bus/FDDI adapter?
/jim
| |||||
| 754.15 | NETRIX::thomas | The Code Warrior | Tue Jan 12 1993 14:25 | 1 | |
2nd generation TC FDDI adapter. Unlike the DEFZA, it's DMA in both directions. | |||||