Peace, Love and C++

I’m fed up with people flaunting how great their machine is because it’s 64 bit. I’ve had 64 bit machines since 2004, when I got a Athlon 64. It’s not new, and it certainly does not mark some “great” revolution in über-processing.

Here are some cool things about 64 bit:

• NX bit
• Fancy new registers
• 64 bit Integers
• More RAM addresses ( bigger address )

You’ll notice, of course, that the NX bit has been built into most modern 32 bit CPUs as well.

PAE ( a sort of intel-esque duct-tape hack to add additional bits to a 32 bit arch ) has been implemented in most modern CPUs. This means that we can up the old RAM cap of 4294967296 bits. PAE lets you address more RAM without going all the way to 64 bit mode.

PAE can also be extended up to 52 bits to address around 4-5 petabytes ( 4.50359963 × 10^15 ).

So, 64 bit users, do you have more then 5 petabytes of RAM? If not, why are you using 64 bit arches? With a 64 bit address, you may end up using more ram then 32 bit mode, so you can store those extra zeros.

I’m going to leave this post off with a qualification: If you’re a developer who works on creating new applications, or working on maintaining old software, please use 64 bit machines.

Since we will have to switch once we go over that 4 petabyte cap, we need to ensure that our code will run on that platform. Just because it sucks right now does not mean it has to suck when we’re forced to migrate.

Endusers: Use 32 bit with PAE.

Hope this puts an end to some bullshit.

kthx.