“AMD signed an agreement with MIPS Technologies Inc. for 64-bit technology, complementing the Alchemy design team already in place at AMD. AMD, Sunnyvale, Calif. said it will license the MIPS64 architecture for an undisclosed period of time, allowing the processor company a growth path for embedded devices. AMD already owns a 32-bit MIPS license through the Alchemy Semiconductor design group AMD acquired in February.” Read the rest of the report at ExtremeTech.
AMD Signs 64-Bit MIPS Processor License
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Will this go to the Hammer or wasted on some PDA??
64 bits is just a tad bit of overkill for a PDA…and Hammer is something else entirely…so maybe they’ll get a contract with SGI? or they could storm the embedded market…those are more likely scenarios
Neat! MIPS laptops mabye? That would be cool.
The 64bit’s will most likly be used for math not bus size. Big registers with vector instructions makes for cheap MPEG (video and audio) playback.
well i guess we now have 64 bit mips, PPC, and x86 (what every you want to call the hammer and intels thing). Guess it’s time to start waiting for 128 bit.
Would a 64-bit X86 compatible be able to emulate a 32bit processor of any type? Its about time we got the ability to run PPC applications on any machine, like we’ve been able to run pentium optimized code anywhere for a while now.
We don’t need more than 64 bit paths or even 128, development goes to the wrong direction! We need to offload main CPUs, so we need more complex ASICs on boards to do specific jobs and we should combine PDA tech with mainframe or desktop solutions. Imagine that you dinamycaly load and program execution units and ACUS if i need image processing or word processing then i’ll load and optimize that code and execute it… future is in the DynMEASU.
Sorry to break it to you, but we do need 64 bit CPUs. The 32-bit address space is incredibly constraining (especially in the kernel). Take systems with large system memories, for example. Currently, you can buy a couple of gigs of RAM for absurdly low prices. In another year or so, you’re going to start hitting that 4GB limit. Second, with RAMs bigger than about 1GB, the OS can’t map all of system memory at the same time, it has to map in portions as needed. For example, say you have some RAM caching part of a file. An app does a read, and the OS determines that the part of the file the app just read is in the cache. Instead of just following the pointer to the page from the cache data structure, it has to first make sure that the page is mapped into kernel RAM, because on > 1GB systems, cache is usually mapped dynamically into the 1GB address space. Then, with 4GB systems, it becomes difficult to use non position independent code. With a 64 bit address space, you can just allocate a range of a few hundred gigs for shared libraries, and you’ll never have collisions.
I clicked Accept, and after a warning saying this would erase the previous data on my hard drive, the real installation took off. I was presented with some commercial information from SuSE while the files were copied.
Yes the article is a sublime detailed technical review