Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Macbooks and CAS 4 RAM

There is no indication to this point from Apple whether or not the EFI based Intel motherboards will automatically adjust latency settings. Since there is no BIOS, there is no way to do it manually. Since before the PowerPC days, Macs have never exploited lower latency.

I would say the possibility of a speed improvement is slim, but I can't say definitively until Barefeats or someone does some bench testing with it.
 
Even if they would use CAS 4 the speed improvement would be slim. Back in the days I spent a lot more for CAS 2 333MHz DDR 512MB and I would never do that again. I would choose a lower CAS if the price difference where slim but I would never hunt for the lowest I could get =P

But now 800 MHz SO-DIMM existed? Have anyone tried them and seen if the machine uses them or if it stays 667MHz? That would make a bigger difference than CAS 4.

Edit: googled:
http://news.softpedia.com/news/DDR2-800Mhz-Memories-for-Notebooks-46906.shtml
I don't know how reliable this is:
https://forums.macrumors.com/posts/3731248/
 
No, we know for sure that 800 MHz would not help at all because the memory controller is hardwired at 667 MHz

Back in the days I spent a lot more for CAS 2 333MHz DDR 512MB and I would never do that again.
Click to expand...

If this was in an Apple machine it's not at all surprising, as I don't believe Macs can take advantage of lower than standard latency.

If the motherboard of a Windows machine can be adjusted and be stable at CL 2.0 vs 2.5 (DDR) or CL4 vs CL5 (DDR2) then there is a theoretical 20% increase in memory bandwidth. Overall realworld speed improvement depends a lot on how the OS and applications use RAM as opposed to all the other bottlenecks in the machine, but I would guess in about the 3% - 4% range.
 
CanadaRAM said:
No, we know for sure that 800 MHz would not help at all because the memory controller is hardwired at 667 MHz



If this was in an Apple machine it's not at all surprising, as I don't believe Macs can take advantage of lower than standard latency.

If the motherboard of a Windows machine can be adjusted and be stable at CL 2.0 vs 2.5 (DDR) or CL4 vs CL5 (DDR2) then there is a theoretical 20% increase in memory bandwidth. Overall realworld speed improvement depends a lot on how the OS and applications use RAM as opposed to all the other bottlenecks in the machine, but I would guess in about the 3% - 4% range.
Click to expand...

I am so proud to be Canadian. CanadaRAM, you are an encyclopedia of RAMformation :D
 
CanadaRAM said:
No, we know for sure that 800 MHz would not help at all because the memory controller is hardwired at 667 MHz
Click to expand...
Are you 100% positive on that? I though that it was only that they didn't existed earlier?

http://www.hothardware.com/Articles/Intel_P965_Showdown__Abit_vs_GIGABYTE/
"While the 975X officially supported DDR2-667 initially (later revisions added DDR2-800 support), the P965's memory controller was designed to handle DDR2 memory running at speeds up to 800MHz out of the gate. This gives the P965 a 2.1GB/sec theoretical memory bandwidth advantage over the 975X when the latter is running at DDR2-667 speeds. DDR2-800 gives the P965 a theoretical peak memory bandwidth of 12.8GB/sec."

For instance? But maybe not for laptops, what do I know. What good does the cpu FSB speed do if the memory can't deliver anyway? Less bandwidth wasted for disk io and similair?
CanadaRAM said:
If this was in an Apple machine it's not at all surprising, as I don't believe Macs can take advantage of lower than standard latency.

If the motherboard of a Windows machine can be adjusted and be stable at CL 2.0 vs 2.5 (DDR) or CL4 vs CL5 (DDR2) then there is a theoretical 20% increase in memory bandwidth. Overall realworld speed improvement depends a lot on how the OS and applications use RAM as opposed to all the other bottlenecks in the machine, but I would guess in about the 3% - 4% range.
Click to expand...
It was in a PC, but afaik it shouldn't make a lot of a difference anyway? So it's that way cl works? I doubt I got 20% performance improvement but would have been sweet if I did :) Or do you mean if one overclock aswell?
3-4% seems more like what I might had got, not worth paying like three times as much for ;D
 
aliquis- said:
Are you 100% positive on that? I though that it was only that they didn't existed earlier?

http://www.hothardware.com/Articles/Intel_P965_Showdown__Abit_vs_GIGABYTE/
"While the 975X officially supported DDR2-667 initially (later revisions added DDR2-800 support), the P965's memory controller was designed to handle DDR2 memory running at speeds up to 800MHz out of the gate. This gives the P965 a 2.1GB/sec theoretical memory bandwidth advantage over the 975X when the latter is running at DDR2-667 speeds. DDR2-800 gives the P965 a theoretical peak memory bandwidth of 12.8GB/sec."

For instance? But maybe not for laptops, what do I know. What good does the cpu FSB speed do if the memory can't deliver anyway? Less bandwidth wasted for disk io and similair?It was in a PC, but afaik it shouldn't make a lot of a difference anyway? So it's that way cl works? I doubt I got 20% performance improvement but would have been sweet if I did :) Or do you mean if one overclock aswell?
3-4% seems more like what I might had got, not worth paying like three times as much for ;D
Click to expand...

Expect more to get 3 fps extra in a game than anything else. For DDR 2 its more trying to bring DDR2 to better than DDR1 speeds as DDR1 still runs games faster than DDR2.
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back