Slashdot Mirror
Sun Open-Sourcing UltraSPARC Design
AKAImBatman writes "While everyone was busy with the holiday season, Sun Microsystems quietly announced the start of the OpenSPARC project. Unlike previous CPUs that were based on the "Open" SPARC specifications (such as LEON), Sun is releasing the complete Verilog source code to their latest and greatest microprocessor. Their current time frame for releasing the source code to the public is in March of 2006. Given their success with the OpenSolaris project, it seems that this is likely to be more than just vaporware. So get out your Virtex FPGAs and your Verilog compilers, and let's get ready to hack some hardware!"
Go to http://www.opencores.org/ for more examples of Open Hardware.
And wasn't UltraSPARC one of the platforms OpenBSD was having difficulty porting to?
:-)
As I understand it, the processor was never the problem. The SPARC architecture is well documented and easily obtainable. It was all the other fiddley bits of hardware that have made life difficult for OSS developers.
Javascript + Nintendo DSi = DSiCade
The first sentence states:
"Sun Microsystems Inc. is looking to ramp up interest in its new UltraSPARC T1 processor by open-sourcing parts of the multicore chip."
For those looking to actually burn an UltraSparc onto their favorite FPGA board are going to be out of luck. Sun couldn't release all the code because they probably have some patents or license agreements.
How is this something new? SPARC has always been, more or less, an open processor design.
Go to http://sparc.org to see.
SPARC already has multiple manufacturers building independent but compatible chips. SPARC was designed to be an open, multi-sourced processor design. Scalable PRocessor ARChitecture.
Tired of FB/Google censorship? Visit UNCENSORED!
http://www.leox.org/docs/faq_MLleon.html
So it looks like Sparc has already been cloned. It's an older version of Sparc, but the one time license fee makes it look as if they never really tried to close the architecture, unlike everyone else.
Mind you, an FPGA Sparc will run at a fairly low clock rate compared to a custom chip, so it's not ideal for a desktop application. And for embedded stuff, I'd suspect that the code density will be too low and the core size too high compared to Arm. And if you can afford to make the custom chips, an Arm license is probably not too bad.
If you're doing an FPGA, the vendor has a bunch of royalty free cores, which are hopefully fairly optimised for the design.
echo -e 'global _start\n _start:\n mov eax, 2\n int 80h\n jmp _start' > a.asm; nasm a.asm -f elf; ld a.o -o a;
check out src.zip in your favorite java dist.
> Personally I enjoy watching Red Hat, Novell/SUSE, Dell and IBM all squirm as Sun
> undercuts their prices in every product line.
And how exactly are they doing this?
> I can get Solaris for free, Sun Cluster for free, the tools for free, Java for free, the
> source code to Solaris for free and a dual core Opteron or multi-core UltraSparc for dirt
> cheap.
So? RHEL is a support contract. I doubt Sun is handing out service contracts for free or even price matching RH. If you want the RH software sans support pick your RHEL Rebuild and start installing. Same for the RH GFS, it is Free as in GPL. Java on the other hand is NOT Free. Sun hardware is getting competitive, which is a good thing but 'dirt cheap'? Put down the crackpipe.
Democrat delenda est
This article refers to UltraSPARC.
Pardon my threadjack, but I just realized that the editors secretly switched my link for a competing brand. Unlike Folgers, I'm afraid it's much cooler to get processor news straight from the horse's mouth.
Javascript + Nintendo DSi = DSiCade
Sun's microSPARC processor has been available for download for quite some time. It is available as synthesizable verilog source code and I think it comes with a PCI master. If sparc is not your style, download their picoJava processor instead.
Try fitting a P4 or UltraSPARC in a Virtex4-1xxLX, you are going to run into several problems.
1) The ASIC runs at 1GHz+ frequency, the V4 implementation would run around 300MHz at best and cost over $10k for the FPGA alone.
2) Most FPGAs block-RAM and LUT-based RAM can be dual-ported at most, this is problematic for register files where a dozen registers may be concurrently accessed during any given cycle. This would require either register duplication or time-multiplexed register access and a corresponding down-clocking of everything else.
3) Logic is expended pretty fast if you do stuff like 64x64 multipliers using logic only. Sure, there are dedicated multipliers in most modern low-cost FPGAs but these are hard-wired to handle DSP-centric MAC operations.
4) People are upset with desktop CPU's power usage but building similar CPUs on FPGAs would require many times more power to achieve the same performance since FPGA's switch fabric and general-purpose programmable elements have way more parasitic capacitance than ASICs' internal hard-wired traces and circuits. With ASIC, 1M logic gates is only ~6M transistors but a ~1M gate-equivalent FPGA with switch fabric and configuration bits goes beyond 50M transistors with much longer routing delays.
FPGAs are not particularly suitable for general-purpose processing where the system has extensive subsystem interdependencies and shared elements. Where they can truly shine is in applications where the data flow is mostly regular and where processing can be broken down into well-defined self-contained stages like telecom, crypto and DSP. Another area where FPGAs can shine is hard-realtime where they can have dedicated logic to handle time-critical events with 100% deterministic deadlines, unlike modern CPUs and OSes where realtime applications have to put up with unpredictable branch mispredicts, cache misses, preemption, out-of-order execution, etc.
That said, the UltraSPARC's verilog source should make for really interesting reading for logic and digital system engineers and academics like myself. This move makes a lot of sense: CPU designers need to hire new talent and this new talent needs to learn about common practice in real-world designs to be of any use or they'll spend most of their first months just catching up. With a real-world design in the wild, CPU-designer job postings could ask people to specify which architectural components they would like to improve and the interviews could steer towards presenting those improvements instead of often irrelevant technicalities.
OpenSolaris has an OSI license. It is called the CDDL. Welcome to open source.
Sun offers the Sun ONE Studio tools for free. Vastly superior to GCC in every measurable way. Of course that is my opinion based on years of code crunching. The fact is that these are available for free. Download and go.
I believe that the source is being made open also.
Absolutely. All of the components under the CDDL are open. Have fun.
More on the way.
Heck, Sun just spent FIVE years working on an entirely new filesystem called ZFS and they released it and open sourced it at the same time. How cool is that?
See : http://www.sun.com/smi/Press/sunflash/2005-11/sunf lash.20051130.1.html
I have heard that .. somewhere. I think Sun does that too. So does my corner store.
see : http://www.redhat.com/sundown/ .. I bought one because it was five times cheaper than my daily coffee intake and I can't live with that either.
Why is there an IBM logo on that page? Why is there an edition RHEL for POWER but not for Sparc ? Why does it say in big BOLD graphics there "Migrate to Linux with IBM + Red Hat"?
Now go look at : http://www.redhat.com/en_us/USA/rhel/compare/serve r/
The absolute cheapest edition is $349 and the top is $2499 !!
I can get Solaris for FREE.
For UltraSparc or for Intel or AMD Opteron.
The cost of an OPTIONAL software support contract is less than 34 cents a day.
I ought to know
See my blog : http://www.blastwave.org/dclarke/blog/pivot/entry. php?id=107
While you're surfing, look at the three guys at :
http://www.novell.com/linux/unixtolinux/
They are all parked on a bench outside the IT Directors office waiting to tell how reiserfs screwed up their data again and they lost the corporate database because of some messed up kernel patch.
But that is just me guessing.
Sure. I agree with "cheap".
Show me a 64-bit Opteron that is faster, cooler and less costly than a SunFire X2100.
Really. Anyone can make junk that is cheap and monsters that are massively expensive.
Show me a 64-bit machine that has more horsepower than an 8-core 1.2GHz SunFire T1000 or a 64-bit AMD Opteron machine with more horsepower than the SunFire X2100.
For less money.
Oh, and the Opteron gear has to be certified to run Windows as well as Linux as well as a real UNIX.
Good luck.
Take a long hard stare at my blog from a little while ago :
http://www.blastwave.org/dclarke/blog/pivot/entry. php?id=113
I count, what? 16 e
Niagra is a real leapfrog forwards, though, assuming it performs in people's real world applications environments as fast as it does in benchmarks. For workloads which are thread partitionable (large numbers of parallel processes, like apache, a java web applications server, etc) it acts much closer to a SMP multiprocessor server with something like 32 cores than either an Intel hyperthreading or HT/multicore, or AMD multicore CPU.
See the T-1000 benchmarks page.
Paraphrasing from that...
The single CPU 1 GHz 8-core T1000 system hs about 3x faster on SPECweb than dual 3.8 GHz Xeons, 2x as fast on SPECjbb business apps benchmarks than dual Xeons, etc
Your typical FPS game will vary, of course, until Carmack gets around to massively multithreading...