[Info-vax] Performance benchmarks, was: Re: VMS Software Inc. OpenVMS 8.4-1H1 Boots on i4 System

[Paul Sture]

On 2015-03-19, Simon Clubley <clubley at remove_me.eisner.decus.org-Earth.UFP>
wrote:
> On 2015-03-18, Michael Moroney <moroney at world.std.spaamtrap.com> wrote:
>> "Robert A. Brooks" <FIRST.LAST at vmssoftware.com> writes:
>>
>>>I don't think there will be much interest by anyone at VSI in playing 
>>>around with anything related to the VAX in the near- or mid-term, as we 
>>>have enough stuff on our collective plates.
>>
>> I don't know.  I entertained a few with a Raspberry Pi emulating a VAX 
>> 3900.
>
> Are there any currently accepted (and freely available) tools for
> benchmarking the performance of VMS systems ? It would be interesting
> to see relative performance both on physical hardware and emulated
> hardware (such as the above as well as the various Alpha emulators).
>
> I'm not looking for something which is industrial strength, but rather
> a tool to get a quick feeling for relative performance between various
> hardware options.

For file I/O, here's the relevant bit of a post by Hoff about David
Mathog's benchmark code from 2001.  As indicated,e network I/O could be an
interesting avenue to explore as well.

> From: Stephen Hoffman <seaohveh at hoffmanlabs.invalid>
> Newsgroups: comp.os.vms
> Subject: Re: I/O Performance (was: Re: on the rocks)
> Date: Fri, 19 Dec 2014 16:28:33 -0500
> 
> If folks here are interested in testing disk I/O performance on various 
> configurations, David Mathog posted benchmark tools and some related 
> data and discussions here in the comp.os.vms newsgroup a decade or more 
> back, and the associated benchmark code is still available at
> 
> <ftp://saf.bio.caltech.edu/pub/software/benchmarks/mybenchmark.zip>  
> 
> There were also more recent discussions of network I/O performance, and 
> Eric Johnson was finding that VMS network performance rather lower than 
> what Linux offered, too.
> 
> Yes, part of the slower disk I/O on VMS is due to the careful nature of 
> disk I/O caching.  Yes, there are ways to turn off parts of that.  Yes, 
> VMS has gotten better over the years, but I'd still assume that VMS I/O 
> performance was slower than that of various other platforms.
> 
> <http://en.wikipedia.org/wiki/PCI_Express>
> 


-- 
Real Daleks don't climb the stairs - real Daleks level the building.