Path: csiph.com!news.mixmin.net!news.unit0.net!peer02.am4!peer.am4.highwinds-media.com!peer02.fr7!futter-mich.highwinds-media.com!news.highwinds-media.com!fx34.am4.POSTED!not-for-mail
Subject: Re: A new benchmark
Newsgroups: comp.lang.c
References: <p4cjpp$g2c$1@solani.org>
From: bartc <bc@freeuk.com>
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:52.0) Gecko/20100101 Thunderbird/52.5.2
MIME-Version: 1.0
In-Reply-To: <p4cjpp$g2c$1@solani.org>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-GB
Content-Transfer-Encoding: 7bit
Lines: 57
Message-ID: <2jnaC.485110$eO2.146871@fx34.am4>
X-Complaints-To: http://netreport.virginmedia.com
NNTP-Posting-Date: Thu, 25 Jan 2018 16:09:34 UTC
Organization: virginmedia.com
Date: Thu, 25 Jan 2018 16:09:48 +0000
X-Received-Body-CRC: 2088703745
X-Received-Bytes: 2503
Xref: csiph.com comp.lang.c:125764

On 25/01/2018 12:50, Philipp Klaus Krause wrote:
> For benchmarking C implementations, the there are a few benchmarks, but
> they all have their problems. Many benchmarks have memory requirements
> that are far too high or need functionality not necessarily available.
> Some are quite one-sided in what they measure (e.g. Whetstone,
> Dhrystone, Coremark).

> 
> I am looking forward to comments from you on this work.
> 
> http://stdcbench.org/

I tried this on my set of 64-bit Windows C compilers (excluding MSVC 
which doesn't work at the minute, and clang which no longer works as it 
piggybacks onto MSVC).

I got:        Optimisation
               Off      On

Pelles C      4133     5920
Lccwin        3341     4954
DMC           4461     5703 (32-bits)
gcc/tdm       3899    12009
Tiny C        3482     3519 (no optimiser)
Mcc           1739     1770 (no optimiser)

(That last is my own C compiler running a temporary code generator; the 
faster one had too many bugs.)

I assume a bigger number is better (otherwise mine wins...).

Building the project however was a small stumbling block as my Make 
program no longer seems to work.

But it can done easily enough by compiling these .c files:

c90base-compression.c
c90base-data.c
c90base-huffman-iterative.c
c90base-huffman-recursive.c
c90base-immul.c
c90base-isort.c
c90base.c
c90double.c
c90float.c
c90struct.c
huffman_tree.c
stdcbench.c
portme.c    (gcc version, & matching .h, from examples directory)

Then linking the corresponding object files into the executable.

But, I've no idea what it is actually testing as it doesn't report much 
apart from that number.

-- 
bartc