AMD Shanghai Opteron: Linux vs. OpenSolaris Benchmarks

[kraftman]

And two other OpenSolaris tests in the old Phoronix benchmark, Linux vs FreeBSD vs OpenSolaris, gained 80% and 100% going from 32bits to 64bits. It seems that the OpenSolaris benches should be increased from 12% to 100%. Then it would look different, right?

Can you show me the results?

I dont get it. What is wrong with the Linux people? Someone posted links to a comparison on a 800MHz SPARC vs 2.8GHz Xeon Linux and claimed that as evidence of Linux being faster. And now above, this moron claims Linux scales very well, and it is just FUD and at the same time he brags about 8CPUs? And this phoronix benchmarks, 32bits against 64bits with an old compiler - how dumb is that? I really really wonder, how can the Linux people call this fair?

I wonder why someone of you don't do benchmarks of Linux and Solaris itself on same machine, using same compilers etc? Who said it was fair benchmark? I bet that Linux scales perfect on much greater number of CPUs. You just saw benchmarks made by "Linux unfriendly and maybe Sun friendly companies", they probably don't even know about problems with crappy malloc glibc library. Maybe just morons claim that Linux doesn't scales well?

And if Solaris scales better why people prefer using Linux?

[Kano]

Some of the benchmarks are just are a bit pointless. The time to compile an app can be used to compare hardware, like different hd, cpu whatever but on SAME os and best only one part changed. But it should not be used between os as different compilers - and even different archs can show different results which you can not simply compared. On 64 bit systems compilers usually work a bit faster, but that's nothing you should worry about. Using compile benchmarks between os is like comparing apples to pears.

[Farrell]

I've been a SunOS/Solaris admin for a long time...and it's been a well know fact that Solaris is not a speed optimized OS. I've always surprised people when I put Linux on a Sun box with how much faster it seems to run. So if Linux is faster on Sun hardware, it is no wonder that Solaris flavours on other platforms are just as slow.

ttyl
Farrell

[etacarinae]

Heh...again, this is about OpenSolaris, not Solaris. They differ a lot nowadays.

I've been a SunOS/Solaris admin for a long time...and it's been a well know fact that Solaris is not a speed optimized OS. I've always surprised people when I put Linux on a Sun box with how much faster it seems to run. So if Linux is faster on Sun hardware, it is no wonder that Solaris flavours on other platforms are just as slow.

ttyl
Farrell

[Farrell]

Heh...again, this is about OpenSolaris, not Solaris. They differ a lot nowadays.

But come from the same code base...and thus the slowness. Unless you are trying to tell me that all of the Solaris code has been written out of OpenSolaris?

I don't think so.

ttyl
Farrell

[kebabbert]

Can you show me the results?

http://www.phoronix.com/forums/showthread.php?t=14017&page=5

"It is possible some of the performance differences are due to the
gcc version being used as Solaris bundles gcc 3.4.3 and other distros
may bundle 4.x, but it is much more likely due to the default ABI
used by the bundled gcc. "gcc -O" on Solaris will default to ia32/x87,
whereas on the other "64 bit" distros tested it will default to amd64.
The performance difference can be seen in the two Byte Computational
benchmarks on page 7 where Solaris appears to lag: Dhrystone 2
(./Run dhry2) and Floating-Point Arithmetic (./Run float). These
tests are compiled with "gcc -O" which produces ia32/x87 code.
When adding "-m64" which puts Solaris on par with the other distros,
the performance jumps quite a bit. Measured on a Intel QX6700,
dhry2 goes from 9307771.4 to 13421763.5 and float goes from
707932.5 to 1477185.6.

The ABI used can make a big difference. Solaris allows you to
choose either, but the default for the bundled gcc is still the
slower ia32/x87."

I wonder why someone of you don't do benchmarks of Linux and Solaris itself on same machine, using same compilers etc? Who said it was fair benchmark? I bet that Linux scales perfect on much greater number of CPUs. You just saw benchmarks made by "Linux unfriendly and maybe Sun friendly companies", they probably don't even know about problems with crappy malloc glibc library. Maybe just morons claim that Linux doesn't scales well?

I doubt that Linux scales well. Ive seen numerous articles where companies say that when their workload increases beyond a certain point, Linux doesnt cut it anymore. And then they switch to Solaris on the same hardware, and everything is fine. Ive posted several links showing this in my posts, on my link above.

And, in your link that moron brags about Linux v2.6 being able to handle 16 cpus. I assume he talks about standard linux kernel. At the same time he talks about large Linux clusters, but fail to tell they are modified Linux kernels.

And if Solaris scales better why people prefer using Linux?

If Solaris scales well on hundreds of CPUs and many more threads, it is not something that affects an ordinary user, right? So normal people which are using dual cpus or so, doesnt notice bad Linux scaling. It is in companies, when their workload increases they have to switch to a real Unix. In my link above, I post an interview to Linux kernel hacker Andrew Morton, who claims that Linux code is full of bugs and regressions. Read the interview if you dont believe me.

And if Linux scales better than Windows, why people prefer using Windows?

[etacarinae]

Once again, they're quite divergent. Only some selected features and enhancements are backported to Solaris from OpenSolaris. And if you think that it takes rewriting "all of the [..] code" to speed up someting than I don't know how to respond to that. It would just mean that any given project can never be optimized.

But come from the same code base...and thus the slowness. Unless you are trying to tell me that all of the Solaris code has been written out of OpenSolaris?

I don't think so.

ttyl
Farrell

[Farrell]

Once again, they're quite divergent. Only some selected features and enhancements are backported to Solaris from OpenSolaris. And if you think that it takes rewriting "all of the [..] code" to speed up someting than I don't know how to respond to that. It would just mean that any given project can never be optimized.

Well, all I know is having used Solaris (and before that, SunOS) and Linux on Sun hardware, the system is almost always faster running Linux, be it SparcStation 10/20 (some with Ross processors), and Ultra/Enterprise beasties of various types. Thus Solaris's offspring, OpenSolaris being slow on x86/64 systems is no surprise to me.

The don't call it Slowaris for nothing!

ttyl
Farrell

[etacarinae]

Sheesh, have you tried OpenSolaris at all ?

Well, all I know is having used Solaris (and before that, SunOS) and Linux on Sun hardware, the system is almost always faster running Linux, be it SparcStation 10/20 (some with Ross processors), and Ultra/Enterprise beasties of various types. Thus Solaris's offspring, OpenSolaris being slow on x86/64 systems is no surprise to me.

The don't call it Slowaris for nothing!

ttyl
Farrell

[yogi_berra]

Virtuozzo has been used for years to create container-style virtulaization for web serving companies. The open source part is OpenVZ.

Plus there is Vserver that also does container-style virtualization.

Not to also mention other virtualization options for Linux... For example Linux now has native built-in ability to be a hypervisor (like Xen or Vmware) in the form of KVM. Every Linux distribution ships with that. There are then a half a dozen virtualization options with various plusses and minuses for Linux.

I think you've missed the point of zones. Zones aren't about virtualization they are about limiting access to a CPU's cores. This comes in extremely handy with Oracle that has a per thread license allowing it to be installed on say an 8 core machine but at a 1 or 2 core price (assuming you limit the Oracle zone to 1 or 2 cores). Container virtualization doesn't limit the access to the CPU's cores thus you would have to pay far more for the Oracle license.

That will upset your CFO which in turn will upset the CEO and in this market probably cost you your job