Date: Mon, 18 Nov 1996 16:53:32 -0500 From: Adrian Antipa <aantipa@certicom.ca> To: FreeBSD-gnats-submit@freebsd.org Subject: misc/2054: Timing benchmarks in Beta Toolkit. Message-ID: <199611182153.QAA09435@galois.certicom.ca> Resent-Message-ID: <199611182150.NAA14389@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
>Number: 2054 >Category: misc >Synopsis: Inaccurate timing benchmarks for ECSVEP2 phase 2. >Confidential: yes >Severity: non-critical >Priority: medium >Responsible: freebsd-bugs >State: open >Class: support >Submitter-Id: current-users >Arrival-Date: Mon Nov 18 13:50:03 PST 1996 >Last-Modified: >Originator: Adrian Antipa >Organization: Certicom Corp. >Release: ECTK-Beta >Environment: DOS 6.20, MSVC 1.5 >Description: eca_mult_affine_point() implements the windowing algorithm. A table is build for each input. It was declared static and a control was set to test the input point. If the input point is the same as in the previous call, then the same table is used, if not, a new table is built. Calling ecs_SharedValueProtocol2_P2() with the same parameters each time, caused a caching of the table construction. For the field size 155, the timing average of ECSVEP2 phase 2 was 1 ms, instead of 120 ms, the timing result for only one loop. The Toolkit binder registered 1 ms for ECSVEP2 for 155 and 239. >How-To-Repeat: Run bsb.exe, the source being /certicom/projects/toolkit/src/bench/bench.c. >Fix: Run only one loop when benchmarking the ECSVEP2. >Audit-Trail: >Unformatted:
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199611182153.QAA09435>