Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Stromasys announces Sparc Virtualization.
49 views
Skip to first unread message
sandy...@mindspring.com
Apr 7, 2015, 10:26:33 AM4/7/15
to
Andrew Gabriel
Apr 7, 2015, 10:39:33 AM4/7/15
to
In article <9799401a-0b9f-4090...@googlegroups.com>,
sandy...@mindspring.com writes:
> http://www.businesswire.com/news/home/20150407005027/en/Stromasys-Brings-Life-Sun-SPARC-applications-CHARON-SSP#.VSPo0I7F9XF
Out of curiosity, is this based on the earlier Transitive product?
--
Andrew Gabriel
[email address is not usable -- followup in the newsgroup]
sandy...@mindspring.com writes:
> http://www.businesswire.com/news/home/20150407005027/en/Stromasys-Brings-Life-Sun-SPARC-applications-CHARON-SSP#.VSPo0I7F9XF
Out of curiosity, is this based on the earlier Transitive product?
--
Andrew Gabriel
[email address is not usable -- followup in the newsgroup]
John D Groenveld
Apr 7, 2015, 11:01:39 AM4/7/15
to
In article <mg0q5b$ovd$1...@dont-email.me>,
<URL:http://en.wikipedia.org/wiki/Charon_%28software%29>
John
groe...@acm.org
Andrew Gabriel <and...@cucumber.demon.co.uk> wrote:
>Out of curiosity, is this based on the earlier Transitive product?
Possibly distinct.
>Out of curiosity, is this based on the earlier Transitive product?
<URL:http://en.wikipedia.org/wiki/Charon_%28software%29>
John
groe...@acm.org
Andrew Gabriel
Apr 7, 2015, 11:53:43 AM4/7/15
to
In article <mg0rgg$mk70$1...@tr22n12.aset.psu.edu>,
Yes, Transitive is described a little here:
http://en.wikipedia.org/wiki/QuickTransit
It doesn't say so there, but they had QuickTransit for
Solaris/SPARC-to-Solaris/x86 working way back in 2003/2004.
I was part of Sun's Solaris kernel team then, and loosly
involved in supporting Transitive in getting it working.
They got it emulating SPARC on x86 faster than SPARC ran
natively. Sun never sold it in the end though.
Later HP got involved with Transitive, and was very interested
in the SPARC/x86 translation because they sold x86 servers which
specifically supported Solaris x86, and hoped to take over some
of the SPARC customer base. This was before Adobe Acrobat
reader was available on Solaris x86, and HP produced a free
Solaris x86 version of Acrobat reader from the SPARC version
by wrapping it in QuickTransit. It worked very well.
http://en.wikipedia.org/wiki/QuickTransit
It doesn't say so there, but they had QuickTransit for
Solaris/SPARC-to-Solaris/x86 working way back in 2003/2004.
I was part of Sun's Solaris kernel team then, and loosly
involved in supporting Transitive in getting it working.
They got it emulating SPARC on x86 faster than SPARC ran
natively. Sun never sold it in the end though.
Later HP got involved with Transitive, and was very interested
in the SPARC/x86 translation because they sold x86 servers which
specifically supported Solaris x86, and hoped to take over some
of the SPARC customer base. This was before Adobe Acrobat
reader was available on Solaris x86, and HP produced a free
Solaris x86 version of Acrobat reader from the SPARC version
by wrapping it in QuickTransit. It worked very well.
Chris Ridd
Apr 7, 2015, 1:52:12 PM4/7/15
to
On 2015-04-07 15:52:44 +0000, Andrew Gabriel said:
> In article <mg0rgg$mk70$1...@tr22n12.aset.psu.edu>,
> groe...@cse.psu.edu (John D Groenveld) writes:
>> In article <mg0q5b$ovd$1...@dont-email.me>,
>> Andrew Gabriel <and...@cucumber.demon.co.uk> wrote:
>>> Out of curiosity, is this based on the earlier Transitive product?
>>
>> Possibly distinct.
>> <URL:http://en.wikipedia.org/wiki/Charon_%28software%29>
>
> Yes, Transitive is described a little here:
> http://en.wikipedia.org/wiki/QuickTransit
> It doesn't say so there, but they had QuickTransit for
> Solaris/SPARC-to-Solaris/x86 working way back in 2003/2004.
> I was part of Sun's Solaris kernel team then, and loosly
> involved in supporting Transitive in getting it working.
> They got it emulating SPARC on x86 faster than SPARC ran
> natively. Sun never sold it in the end though.
Ouch, that must have been slightly embarrassing!
> In article <mg0rgg$mk70$1...@tr22n12.aset.psu.edu>,
> groe...@cse.psu.edu (John D Groenveld) writes:
>> In article <mg0q5b$ovd$1...@dont-email.me>,
>> Andrew Gabriel <and...@cucumber.demon.co.uk> wrote:
>>> Out of curiosity, is this based on the earlier Transitive product?
>>
>> Possibly distinct.
>> <URL:http://en.wikipedia.org/wiki/Charon_%28software%29>
>
> Yes, Transitive is described a little here:
> http://en.wikipedia.org/wiki/QuickTransit
> It doesn't say so there, but they had QuickTransit for
> Solaris/SPARC-to-Solaris/x86 working way back in 2003/2004.
> I was part of Sun's Solaris kernel team then, and loosly
> involved in supporting Transitive in getting it working.
> They got it emulating SPARC on x86 faster than SPARC ran
> natively. Sun never sold it in the end though.
> Later HP got involved with Transitive, and was very interested
> in the SPARC/x86 translation because they sold x86 servers which
> specifically supported Solaris x86, and hoped to take over some
> of the SPARC customer base. This was before Adobe Acrobat
> reader was available on Solaris x86, and HP produced a free
> Solaris x86 version of Acrobat reader from the SPARC version
> by wrapping it in QuickTransit. It worked very well.
branded it "Rosetta" but it was QuickTransit licensed from Transitive.
--
Chris
Nemo
Apr 7, 2015, 5:57:55 PM4/7/15
to
On 07/04/2015 11:52, Andrew Gabriel wrote:
[...]
[...]
> It doesn't say so there, but they had QuickTransit for
> Solaris/SPARC-to-Solaris/x86 working way back in 2003/2004.
> I was part of Sun's Solaris kernel team then, and loosly
> involved in supporting Transitive in getting it working.
> They got it emulating SPARC on x86 faster than SPARC ran
> natively.
At similar clock speeds? I am somewhat skeptical here.
> Solaris/SPARC-to-Solaris/x86 working way back in 2003/2004.
> I was part of Sun's Solaris kernel team then, and loosly
> involved in supporting Transitive in getting it working.
> They got it emulating SPARC on x86 faster than SPARC ran
> natively.
Andrew Gabriel
Apr 7, 2015, 7:40:38 PM4/7/15
to
In article <ypYUw.155404$tF6....@fx11.iad>,
By 2003/2004, SPARC was falling well behind x86 in performance,
and Sun was losing SPARC customers who needed better performance.
(SPARC was only viable for applications which were too large to
fit on the largest commodity x86 systems of the day, and there
weren't anything like enough of those to keep Sun going.)
That situation continued until Sun started selling the Fujitsu
SPARC64 systems (M4000, M5000, et al) in 2007.
That's partly why Sun did a big x86 push during this time, both
in Solaris and in the hardware teams.
and Sun was losing SPARC customers who needed better performance.
(SPARC was only viable for applications which were too large to
fit on the largest commodity x86 systems of the day, and there
weren't anything like enough of those to keep Sun going.)
That situation continued until Sun started selling the Fujitsu
SPARC64 systems (M4000, M5000, et al) in 2007.
That's partly why Sun did a big x86 push during this time, both
in Solaris and in the hardware teams.
Nemo
Apr 8, 2015, 10:07:43 AM4/8/15
to
On 07/04/2015 19:39, Andrew Gabriel wrote:
> In article<ypYUw.155404$tF6....@fx11.iad>,
> Nemo<ne...@invalid.invalid> writes:
>> On 07/04/2015 11:52, Andrew Gabriel wrote:
>> [...]
>>> It doesn't say so there, but they had QuickTransit for
>>> Solaris/SPARC-to-Solaris/x86 working way back in 2003/2004.
>>> I was part of Sun's Solaris kernel team then, and loosly
>>> involved in supporting Transitive in getting it working.
>>> They got it emulating SPARC on x86 faster than SPARC ran
>>> natively.
>>
>> At similar clock speeds? I am somewhat skeptical here.
>
> By 2003/2004, SPARC was falling well behind x86 in performance,
> and Sun was losing SPARC customers who needed better performance.
> (SPARC was only viable for applications which were too large to
> fit on the largest commodity x86 systems of the day, and there
> weren't anything like enough of those to keep Sun going.)
> That situation continued until Sun started selling the Fujitsu
> SPARC64 systems (M4000, M5000, et al) in 2007.
Granted but that still does not answer my question: What were the clock
> In article<ypYUw.155404$tF6....@fx11.iad>,
> Nemo<ne...@invalid.invalid> writes:
>> On 07/04/2015 11:52, Andrew Gabriel wrote:
>> [...]
>>> It doesn't say so there, but they had QuickTransit for
>>> Solaris/SPARC-to-Solaris/x86 working way back in 2003/2004.
>>> I was part of Sun's Solaris kernel team then, and loosly
>>> involved in supporting Transitive in getting it working.
>>> They got it emulating SPARC on x86 faster than SPARC ran
>>> natively.
>>
>> At similar clock speeds? I am somewhat skeptical here.
>
> By 2003/2004, SPARC was falling well behind x86 in performance,
> and Sun was losing SPARC customers who needed better performance.
> (SPARC was only viable for applications which were too large to
> fit on the largest commodity x86 systems of the day, and there
> weren't anything like enough of those to keep Sun going.)
> That situation continued until Sun started selling the Fujitsu
> SPARC64 systems (M4000, M5000, et al) in 2007.
speeds?
Casper H.S. Dik
Apr 8, 2015, 11:03:41 AM4/8/15
to
Nemo <ne...@invalid.invalid> writes:
>> By 2003/2004, SPARC was falling well behind x86 in performance,
>> and Sun was losing SPARC customers who needed better performance.
>> (SPARC was only viable for applications which were too large to
>> fit on the largest commodity x86 systems of the day, and there
>> weren't anything like enough of those to keep Sun going.)
>> That situation continued until Sun started selling the Fujitsu
>> SPARC64 systems (M4000, M5000, et al) in 2007.
>Granted but that still does not answer my question: What were the clock
>speeds?
Many systems were below 1GHz and even the T1/T2 series were not
>> By 2003/2004, SPARC was falling well behind x86 in performance,
>> and Sun was losing SPARC customers who needed better performance.
>> (SPARC was only viable for applications which were too large to
>> fit on the largest commodity x86 systems of the day, and there
>> weren't anything like enough of those to keep Sun going.)
>> That situation continued until Sun started selling the Fujitsu
>> SPARC64 systems (M4000, M5000, et al) in 2007.
>Granted but that still does not answer my question: What were the clock
>speeds?
much higher; and none of them were out-of-order.
At the same time Opterons run more than 2GHz, OoO and with
bigger bandwith/lower latency memory.
It did need to 64 bit CPUs, of course, and AMD was the only
provider at that time, IIRC.
Casper
0 new messages