Slashdot Mirror

...about the only OS sluggish on modern hardware is Vistrash.

What's your point?

Solaris doesn't have speed issues.

If you use older hardware, Linux has speed issues too.

Try doing a full install of Fedora Core 9 & CentOS5 on an old Celeron 600Mhz system with 512mb of RAM, do that same install on Solaris; no, you don't get to turn any services off or make any speed tunings on either install.

You will find surprise, both systems are painfully slow.

Solaris seems to be faster.

Others have also measured Solaris to be faster.

Granted, you could benchmark Damn Small Linux against Solaris fully decked out, and running ZFS, but that wouldn't be fair.

Nowadays, there is hardly any general difference between Linux and Solaris, when they are properly tuned, it's all about how much software is running on the system, and which system services are running.

I.E. ZFS. Naturally, I am most interested in differences in the default (full) installs, because most *IX users are performing default full installs of popular distros, i.e. Redhat, Mandriva, SuSE.

And not optimized niche distros like dsl.

The Processor is about equivalent to a mid-to-late Pentium III.

No, it's actually about the same as a Powermac G5 at the moment. Read more here.

"Results
Overall Score
PlayStation 3
105.2
Power Mac G5
106.9"

The 3D functions of the GPU are locked...you can't get any 3D acceleration in Linux. More importantly, re: it being a supercomputer, the last I heard it was being beaten in benchmarks by a G5. With code specially optimized for the vector processors, it may be very fast for specialized tasks, but currently its just a crippled PowerPC chip (no out-of-order execution).

Do people read article anymore??

First off the PS3 is a computer, period. Get over it.

Second, from the article:

"Geekbench also isn't able to exploit the eight vector processors on the Cell processor. Any program designed and optimized for the Cell processor should be a lot faster than one designed for a generic processor (like, say, Geekbench). So while the Geekbench results might seem disappointing, keep in mind that Geekbench can't exercise the PlayStation 3 to its full potential."

http://www.geekpatrol.ca/2006/11/playstation-3-per formance/

There are running a stock version of Linux which does not take advantage of the Cell processors capability. Christ it's like running a Intel Processor with HT disabled! What did you expect.

The Cell processor is one hell of a processor, read the technical specs. It's best use may not be in the PS3. But since SONY is not in the OS market, they are left up to 3rd parties to develop an OS to work on the Cell.

My complaint is the lack of memory. I would like the ability to up the memory to 1 GB. Show me a PS3 with 1GB and you'd have a very interesting computing device.

Harrison says that the current PS3 game lineup is using less than half of the machines power, adding that 'nobody will ever use 100 percent of its capacity.'

There are some numbers on MacBook Pro performance here. The bottom line is that the they're 10% faster than the last versions at the same clock speed for 32-bit tasks and about 20% faster for 64-bit things (than the old model was for 32-bit ones, since the Core 1 couldn't do 64-bit).

The geekpatrol blog you referenced in your post did some MacBook Pro Core 2 Duo benchmarking, and they tested in both 32-bit and 64-bit mode. So it seems to me that they support it.

I sent Brian a question (shortly before he got "Slashdotted") about whether the new MacBook Pros supported 64 bit mode. He was kind enough to compile and run the little 'sizeof' program I sent him and respond about the output..

The announcements and marketing information about the new MBP's was conspicuously absent of any mention of the 64 bit support of the new Merom / Core 2 Duo processor. This is strange because both the Mac Pro and the iMac specifically mention their 64 bit support.

There is a lot of contradictory information floating around about the state of 64 bit support on the Intel Macs. So, I asked him to compile an app to show the sizeof a long int and pointer. The output showed 4 Bytes / 32 bits.

So, this is curious... Does x86 Tiger not support 64 bit mode? But, people have done tests on 32 bit vs. 64 bit on intel Macs ( http://www.geekpatrol.ca/blog/150/ ). So, why is the MBP different than the iMac, which uses the same processor and chipset?

Anyone have more definitive information on 64 bit support for this new MacBook Pro? Or for x86 Tiger, the new iMacs and Mac Pro's?

Also, before all the "64 bit support is pointless" replies; yes, I know it can only handle 3GB of RAM. I know the benefits of 64 bit will not be dramatic (I already have two Linux boxes running Athlon64's in 64 bit mode). I'm just curious whether all the features of the processor can be used. I also want the performance benefit of doubling the number of general purpose processors and 64 bit math. And, since Leopard is supposed to have much better 64 bit support, I want to see where this MacBook Pro will stand.

I rarely trust benchmarking because often there are tradeoffs that make one area perform poorly and another well. That said, take a look at this benchmark - it shows some serious problems (stdlib allocate is 3000+% slower!), but also some stuff that macosx does well (blowfish, stdlib write).

http://www.geekpatrol.ca/blog/106/

I expect macosx would trounce XP in 2d graphics tests because quartz extreme makes heavy use of graphics hardware and we won't have that until Vista for Windows. I don't know of any hardware accelerated GUI features for Linux but I'm running the fairly old KDE and badly need to update my box (my distribution is no more and I need to tarball and archive about 200MB of data [mostly text files], then switch distributions).

Personally, I don't think microkernel is bad, but it has the same advantages and shortcomings as a true object oriented languages like Smalltalk. For illustration, in Smalltalk, you could have a case where object2 may need to alter data from object1. Object2 requests the data from object1, object1 sends the data to object2, object2 checks if changes are needed and if so, updates the data and then sends a message with the updated data back to object1. In C++, that could be accomplished simply by making the data public or protected and allowing the other object to directly access its data (though you can also write helper functions that do messaging exactly the same as above). The good thing about messaging is you don't have to worry about another object breaking your data as much and it makes the code much more modular (easier to test and debug, IMO) and reusable.

Which independent benchmarks confirming that the Intel Core Duo really is about 4x the speed of the G4, I'd say the Mini just got a whole lot more viable. At $800 the price is a significant step up, but I guess you gotta pay to play, and it's still the cheapest Duo system I've noticed.