> There were absolutely no applications in that review that
> required 64-bits. Absolutely everything ran perfectly on Core
> Duo.
Fine. Show me a benchmark where you stitch together 50,000 x 50,000 pixel images with 32-bits.
> If you are pounding away on your disk, you are receiving a lot
> of I/O requests, which interrupt your CPU and prevent it from
> running real work going on simultaneously. If you have dual
> cores, however, you can have one CPU service interrupts, while
> the other one computes work, and you absolutely will finish
> faster.
Which has a higher latency: a CPU unterrupt or a disk seek? You
may finish sooner but it's still a lousy way to work.
I do builds on my notebook which pound the heck out of the drives that I use. Two have failed in the last 5 months and
these are high-quality Hitachi drives. If you're pounding
away on your notebook drive with an antivirus scan, you're
going to see a lag on everything else that you're doing related
to disk access. Which is why you do your virus scans when you
aren't using your machine. Or that particular disk.
> Why don't you finish reading the review? You seem to be
> stopping at Winstone and SysMark, while apparently oblivious
> to the 88% improvement in 3DStudio Max, the 87% improvement
> in Adobe Premier, the 36% improvement in Adobe Photoshop, the
> 65% improvement in DVD Shrink, the 64% improvement in WME,
> the 65% improvement in Quicktime encoding, the 64%
> improvement in iTunes, and the 12% improvement in Quake IV.
I did read that and I will give that to you. I think that I mentioned most of those anyways.
> Word and Excel don't need to be accelerated any more than
> they are now. People don't care if your computer can run
> Word or Excel faster, because almost any operation is
> already instantaneous.
Well I do because I work with some pretty big specs. And
checkpointing hits the disks. I've worked in the accounting
world as well and some of these guys come up with some pretty
big spreadsheets. And then there's the really big application:
PowerPoint which can be slow depending on what you're doing.
> But if you can rip a DVD in 2:00 as opposed to 3:20, then
> you are going to find some very interested end-users.
Theoretically, if you're going to rip a DVD, you'd be better
off with x64 dual core than x32 dual core if you did the
programming right. As you could take advantage of the
additional SIMD registers to keep more working data in registers. Were those benchmarks done with 32-bit software or 64-bit software on the X2s?
> Now how about you show me some 64-bit apps that can sport
> these kinds of gains. Go ahead and use 10x more memory, if
> you want, but it won't help.
I know that Firefox will run faster on 64-bits than 32-bits.
And I know that the capability to run faster on 64-bits dc is there compared to 32-bits dc. But someone has to do the programming. Perhaps you can show me a benchmark where 32-bit DC beats 64-bits DC on a rendering benchmark where there is 64-bit software. From an engineering point of view your argument doesn't hold water.
> Now how about you show me some 64-bit apps that can sport
> these kinds of gains. Go ahead and use 10x more memory, if
> you want, but it won't help.
Fine. 80 GB system (Main Memory) RamDisk with Virus Scan and Microsoft Office on a 64-bit system vs a Dual-Core system. Your hard-drive latency will kill the dual-core system.
Market Data 
Markets 
AI
