reisert@ricks.enet.dec.com (Jim Reisert) (08/22/90)
Here are the results of the last set of measurements I have done. They
attempt to measure the following effects:
* P.R. Glassel & Assoc., Inc. QCACHE 4.20
* BUFFERS=4 vs. BUFFERS=16
* when running under Quarterdeck's DESQview
* as measured by Multisoft's PC-Kwik benchmark
NOTE: I attempted to measure the performance of HYPERDRIVE, but the
shareware copy I was to test did not have all the features enabled,
such as write buffering, so I didn't test it.
1. Measurement of P.R. Glassel & Assoc., Inc. QCACHE 4.20
Here is how QCACHE compared to other disk caches as measured with the PC
Labs Benchmark 5.0 (the tests are described at the end). The cache size in
all cases was 384K, and QCACHE performed identically in conventional and
expanded memory:
TEST1 TEST2 TEST3 TEST4
no cache 58.50 8.86 5.44 3.99 secs.
SMARTDRV 35.06 7.26 4.00 3.33 "
PC-CACHE 5.5 30.29 6.70 3.73 3.32 "
PC-CACHE 6.03 13.13 8.53 3.14 2.75 "
QCACHE 4.20 9.87 9.18 3.96 3.65 "
SuperPCK 3.56 8.52 4.94 1.83 2.45 "
2. Effect of BUFFERS
I ran the PC Labs Benchmark tests (described below). The first run had
BUFFERS=4 and the second had BUFFERS=16:
BUFFERS=4 TEST1 TEST2 TEST3 TEST4
no cache 58.50 8.86 5.44 3.99 secs.
SuperPCK 3.56 8.52 4.94 1.83 2.45 "
BUFFERS=16 TEST1 TEST2 TEST3 TEST4
no cache 58.05 8.82 4.92 4.06 secs.
SuperPCK 3.56 8.64 4.91 2.63 2.44 "
The setting of BUFFERS seemed to have no effect, except in TEST3 (see
description below).
3. Running under DESQview
Here are the results of running with three different caches, using a single
512K DOS window in DESQview. The results of running with cache in expanded
or extended memory are the same:
TEST1 TEST2 TEST3 TEST4
no cache 69.79 12.05 10.95 10.63 secs.
PC-CACHE 5.5 31.61 7.61 5.13 4.56 "
PC-CACHE 6.03 16.05 12.24 10.47 10.24 "
SuperPCK 3.56 32.54 8.26 5.32 4.97 "
For comparison, here are the same tests, run in stand-alone mode:
TEST1 TEST2 TEST3 TEST4
no cache 58.50 8.86 5.44 3.99 secs.
PC-CACHE 5.5 30.29 6.70 3.73 3.32 "
PC-CACHE 6.03 13.13 8.53 3.14 2.75 "
SuperPCK 3.56 8.52 4.94 1.83 2.45 "
4. PC-Kwik Benchmark
Multisoft provides a disk cache benchmarking package with their Power Pack
(Does it come with Super PC-Kwik also?). Their benchmark runs two tests:
512 records of 512 bytes, and 64 records of 4096 bytes.
Here are the results of using a 384K disk cache in expanded memory:
TEST1 TEST2
no cache 54.10 8.79 secs.
SMARTDRV 31.75 6.87 "
PC-CACHE 5.5 29.39 5.82 "
SuperPCK 3.56 13.46 4.94 "
PC-CACHE 6.03 3.68 6.59 "
Here are the results of using a 384K disk cache in extended memory:
TEST1 TEST2
no cache 54.10 8.79 secs.
SMARTDRV 30.65 6.76 "
PC-CACHE 5.5 29.49 5.88 "
SuperPCK 3.56 11.50 6.18 "
PC-CACHE 6.03 13.51 5.99
The results are almost identical, except PC-CACHE 6.03 and SuperPCK 3.56
both seemed to perform worse in extended memory.
I think I have satisfied everyone's requests for measurements. If not,
please let me know.
jim
DESCRIPTION of SYSTEM and TESTS:
All tests were run on a 20 MHz 386 machine with a 42 MB hard disk (that was
defragmented at the end of each test) and 2 MB RAM running MS-DOS 4.01. I
used QEMM as my EXPAnded memory manager. All tests were run at the DOS
prompt (as opposed to under Windows or DESQview), unless otherwise noted.
The following tests comprise the PC-Labs Benchmark Test 5.0 suite:
TEST1 - 512 records of 512 bytes each
TEST2 - 64 records of 4096 bytes each
TEST3 - 16 records of 16384 bytes each
TEST4 - 8 records of 32768 bytes each
Each test had 5 stages: 1) file creation, 2) sequential write, 3) sequential
read, 4) random write and 5) random read. The results are listed in
seconds.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
"The opinions expressed here in no way represent the views of Digital
Equipment Corporation."
James J. Reisert Internet: reisert@ricks.enet.dec.com
Digital Equipment Corp. UUCP: ...decwrl!ricks.enet!reisert
77 Reed Road
Hudson, MA 01749-2895