We're currently running a 3 GHz Pentium with 2 GB memory under Windows 2000. We hope to speed things up by 15-20%, by going to AMD X86-64 and / or Linux. Has anybody tried this? Any feedback?
Need to speed up Stratix compiles.
Started by ●February 27, 2004
Reply by ●February 29, 20042004-02-29
No, but I'd like to know the results on the AMD 64. I spead up my 3.2GHz P4 compliles by 20% by making sure my memory was running at dual channel 400MHz. Turned out it was running dual channel 333MHz. I had to actually downgrade my memory slightly, because my motherboard saw cas2 DDR and dropped to 333. With cas 2.5 it was confident to go to 400. YMMV, Ken "Pete Fraser" <pete@rgb.com> wrote in message news:103vj8jovcv761e@news.supernews.com...> We're currently running a 3 GHz Pentium with 2 GB > memory under Windows 2000. > > We hope to speed things up by 15-20%, by going > to AMD X86-64 and / or Linux. > > Has anybody tried this? > Any feedback? > >
Reply by ●March 2, 20042004-03-02
I went from a 2.5GHz Pentium to a 3GHz Xeon and got a very consistent 33% speed increase in Stratix compiles and SOPC Builder generation. I suspect the increased cache size is the most critical thing, since clock rate increased by only 20%, but I'm only speculating. Both machines ran XP and RAM was 1GB in both machines. I'm curious to hear how your compiles improve with AMD/Linux. -- Pete "Pete Fraser" <pete@rgb.com> wrote in message news:<103vj8jovcv761e@news.supernews.com>...> We're currently running a 3 GHz Pentium with 2 GB > memory under Windows 2000. > > We hope to speed things up by 15-20%, by going > to AMD X86-64 and / or Linux. > > Has anybody tried this? > Any feedback?
Reply by ●March 2, 20042004-03-02
Pete Fraser wrote:> > We're currently running a 3 GHz Pentium with 2 GB > memory under Windows 2000. > > We hope to speed things up by 15-20%, by going > to AMD X86-64 and / or Linux. > > Has anybody tried this? > Any feedback?I can't tell you what to expect from an AMD64 with today's software, but I expect this will be the platform of choice for the next couple of years. It may not be the best investment at the moment, but I expect by the end of the year, much of the software will be optimized for 64 bit operation and you will see over half the new engineering workstations running an AMD64 processor. IIRC, AMD is producing a low cost version of the AMD64. I expect sales will take off very quickly. Once these start showing up on the software developer's desks we will see them optimizing for it. -- Rick "rickman" Collins rick.collins@XYarius.com Ignore the reply address. To email me use the above address with the XY removed. Arius - A Signal Processing Solutions Company Specializing in DSP and FPGA design URL http://www.arius.com 4 King Ave 301-682-7772 Voice Frederick, MD 21701-3110 301-682-7666 FAX
Reply by ●March 2, 20042004-03-02
On Fri, 27 Feb 2004 15:04:46 -0800, Pete Fraser wrote:>We're currently running a 3 GHz Pentium with 2 GB >memory under Windows 2000.The server versions of Win2000 can handle up to 32GB of RAM, if that's the main limiting factor. Can get pricey, though :o((>We hope to speed things up by 15-20%, by going >to AMD X86-64 and / or Linux. > >Has anybody tried this? >Any feedback?Unless there's a version of the compiler specifically for a 64-bit architecture, then you're unlikely to see any real speed gain to justify the cost, and even if there is, I doubt the gains would be all that impressive. 64-bit CPUs really only come into their own in applications that need to access very large virtual address spaces (>4GB). Mostly, that's server-type apps. The need for 64-bit arithmetic is likely very small in this case. Can the compiler multi-thread? If so, a mobo with a couple of HT Xeons (4 CPUs), will give you all the extra horsepower you'll need. If not, a dual-processor system would still perform a lot better, since one CPU can work flat-out on the compile, while the other is handling the OS and other background tasks. All this assumes that the compiler's performance is, in fact, CPU or memory bound as you imply. Are you actually certain that this is indeed the case? Might a faster disk system help? -- Max
Reply by ●March 2, 20042004-03-02
"Pete Fraser" <pete@rgb.com> writes:> We're currently running a 3 GHz Pentium with 2 GB > memory under Windows 2000. > > We hope to speed things up by 15-20%, by going > to AMD X86-64 and / or Linux. > > Has anybody tried this?Quartus II 3.0 does not run on X86-64. See news:<87k76ox8ya.fsf@zener.home.gustad.com> or http://tinyurl.com/2pvlj The fix should be trivial since it's just the driver script which does not recognize the architecture. If it tried to run some X86 code rather than checking uname it would work. I dunno about 4.0 though. Anybody tried? Petter -- A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing? A: Top-posting. Q: What is the most annoying thing on usenet and in e-mail?
Reply by ●March 2, 20042004-03-02
Seems to be a common misconception that 64bits just increases the amount of addressable memory. More importantly for most applications is that twice the data is moved or operated on per clock cycle. Ken "Max" <mtj2@btopenworld.com> wrote in message news:6ig9409fn2h2futm3hjjmkd4uiv4lpkmu2@4ax.com...> On Fri, 27 Feb 2004 15:04:46 -0800, Pete Fraser wrote: > > >We're currently running a 3 GHz Pentium with 2 GB > >memory under Windows 2000. > > The server versions of Win2000 can handle up to 32GB of RAM, if that's > the main limiting factor. Can get pricey, though :o(( > > >We hope to speed things up by 15-20%, by going > >to AMD X86-64 and / or Linux. > > > >Has anybody tried this? > >Any feedback? > > Unless there's a version of the compiler specifically for a 64-bit > architecture, then you're unlikely to see any real speed gain to > justify the cost, and even if there is, I doubt the gains would be all > that impressive. > > 64-bit CPUs really only come into their own in applications that need > to access very large virtual address spaces (>4GB). Mostly, that's > server-type apps. The need for 64-bit arithmetic is likely very small > in this case. > > Can the compiler multi-thread? If so, a mobo with a couple of HT Xeons > (4 CPUs), will give you all the extra horsepower you'll need. > If not, a dual-processor system would still perform a lot better, > since one CPU can work flat-out on the compile, while the other is > handling the OS and other background tasks. > > All this assumes that the compiler's performance is, in fact, CPU or > memory bound as you imply. Are you actually certain that this is > indeed the case? Might a faster disk system help? > > -- > Max
Reply by ●March 2, 20042004-03-02
On the disk speed issue I have one data point. I upgraded my 1GHz PIII-M laptop drive from a slow 4200 RPM to the fastest 7200 RPM available (for laptops) and my Nios system build went from about 16 min. to about 15 min. Not worth the pain and expense of swapping the drive. On memory, I upgraded the memory in my 3.2 GHz P4 from 512 to 1GB and there was no noticable difference until I set the memory from 333MHz to 400MHz dual channel. Then my system build went from 5 min. to 4 min. - 20%. Ken "Max" <mtj2@btopenworld.com> wrote in message news:6ig9409fn2h2futm3hjjmkd4uiv4lpkmu2@4ax.com...> On Fri, 27 Feb 2004 15:04:46 -0800, Pete Fraser wrote: > > >We're currently running a 3 GHz Pentium with 2 GB > >memory under Windows 2000. > > The server versions of Win2000 can handle up to 32GB of RAM, if that's > the main limiting factor. Can get pricey, though :o(( > > >We hope to speed things up by 15-20%, by going > >to AMD X86-64 and / or Linux. > > > >Has anybody tried this? > >Any feedback? > > Unless there's a version of the compiler specifically for a 64-bit > architecture, then you're unlikely to see any real speed gain to > justify the cost, and even if there is, I doubt the gains would be all > that impressive. > > 64-bit CPUs really only come into their own in applications that need > to access very large virtual address spaces (>4GB). Mostly, that's > server-type apps. The need for 64-bit arithmetic is likely very small > in this case. > > Can the compiler multi-thread? If so, a mobo with a couple of HT Xeons > (4 CPUs), will give you all the extra horsepower you'll need. > If not, a dual-processor system would still perform a lot better, > since one CPU can work flat-out on the compile, while the other is > handling the OS and other background tasks. > > All this assumes that the compiler's performance is, in fact, CPU or > memory bound as you imply. Are you actually certain that this is > indeed the case? Might a faster disk system help? > > -- > Max
Reply by ●March 3, 20042004-03-03
Hi Max,> All this assumes that the compiler's performance is, in fact, CPU or > memory bound as you imply. Are you actually certain that this is > indeed the case? Might a faster disk system help?Provided the peak memory consumption of Quartus for the compilation in question is less than the amount of physical memory in the system, increasing the amount of memory will not help compile time. For non-trivial designs, a Quartus compile will be most heavily influenced by CPU speed, and then by memory sub-system speed -- disk speed will have little influence. CAD tools process a lot of data. I don't know if a Xeon (bigger cache) is much faster than a normal P4 (smaller cache), but I wouldn't be surprised if this were the case for the same reason that a Xeon processor is supposedly better for server applications -- bigger cache helps applications whose data set doesn't fit into the cache. Regards, Paul Leventis Altera Corp.
Reply by ●March 3, 20042004-03-03
> Seems to be a common misconception that 64bits just increases the amountof> addressable memory. More importantly for most applications is that twice > the data is moved or operated on per clock cycle.64-bitness _is_ mostly about addressable memory -- it is rare that 64-bit integers help reduce run-time. Please see my previous postings on the topic and some of the replies to it: http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&oe=UTF-8&q=Paul+Leventis+64-bit Regards, Paul Leventis Altera Corp.