 |
 |
Intel iMac - Diff between 4200 and 5300 DDR2?
|
 |
|
 |
|
Mac Elite
Join Date: Jun 2005
Location: UK
Status:
Offline
|
|
Hi, when my Intel iMac arrives, I am planning to put extra RAM in it, but its damn expensive. Will there be a big performance difference between 5300 DDR2 and 4200 DDR2, as the latter is less expensive, hence I could get an extra 1 GB instead of just 512
Cheers
|
|
iMac Core Duo 1.83 Ghz | 1.25GB RAM | 160HD, MacBook Core Duo 1.83 Ghz | 13.3" | 60HD | 1.0GB RAM
|
| |
|
|
|
 |
|
 |
|
Fresh-Faced Recruit
Join Date: Jan 2006
Status:
Offline
|
|
Buying slower memory will mean the overall system performance being slower because the bus will have to run at the slower memory speed. The bus speed is even more important for dual-core processors.
|
|
|
| |
|
|
|
 |
|
 |
|
Posting Junkie
Join Date: Oct 2001
Location: South of the Mason-Dixon line
Status:
Offline
|
|
The bus speed of the RAM should not change based on the rating of the installed RAM modules. But I won't swear it can't.
You should use *at minimum* the same speed RAM module the unit ships with. Plugging in faster RAM will yield no benefit whatsoever. Slower RAM will likely result in the inability of the machine to boot.
If it ships with DDR 5300, then add modules rated for DDR5300.
|
|
|
| |
|
|
|
 |
|
 |
|
Posting Junkie
Join Date: Oct 2005
Location: Houston, TX
Status:
Offline
|
|
The FSB from the chipset to the CPU will still run at 667Mhz, but the memory bus would be slowed to 533Mhz.
Intel's chipsets generally do support running memory at a variety of lower clockrates (machines that support DDR2-533 can also run DDR2-400), but Apple's computers generally don't, so I can't say if it will work or not.
|
|
|
| |
|
|
|
 |
|
 |
|
Junior Member
Join Date: Apr 2005
Location: Colorado
Status:
Offline
|
|
Xbench Results
These are two Xbenches I ran with and without the DDR2-4200 chip installed. As you can see, the memory performance is affected around 7points overall. Not bad for memory that averages to be 33% cheaper than the fast stuff!
|
|
MacBook Pro (Mid 2007), 2.4Ghz, 2GB DDR2-667Mhz, 160GB, Superdrive, Nvidia Geforce 8600M GT w/256MB, 15.4" WXGA+ LCD
|
| |
|
|
|
 |
|
 |
|
Forum Regular
Join Date: Mar 2005
Status:
Offline
|
|
Originally Posted by fleaplus
Xbench Results
These are two Xbenches I ran with and without the DDR2-4200 chip installed. As you can see, the memory performance is affected around 7points overall. Not bad for memory that averages to be 33% cheaper than the fast stuff!
Where are you seeing a price difference? Memory.com has the 5300 1GB's for $123 and the 4200 1GB's for $120.
I've seen the 5300's as low as $108. Are you saying you found someone selling 1 GB 4200 sticks for around $70?
Of course, for me, I'm tempted to get the 4200 because I can't find the 5300 locally.
-Todd
|
|
The moderators in this forum have too much time on their hands.
|
| |
|
|
|
 |
|
 |
|
Junior Member
Join Date: Apr 2005
Location: Colorado
Status:
Offline
|
|
Originally Posted by toddtmw
Where are you seeing a price difference? Memory.com has the 5300 1GB's for $123 and the 4200 1GB's for $120.
I've seen the 5300's as low as $108. Are you saying you found someone selling 1 GB 4200 sticks for around $70?
Of course, for me, I'm tempted to get the 4200 because I can't find the 5300 locally.
-Todd
I ordered two of these Kingston brand ram for my Dell Inspiron 9300 a while back. Although this GSkill brand ram is probably the most lucrative of them all. The Kingston has work fine for 3 days now, other than the insertion misalignment that cause the machine to not boot until I fixed it (which could happen with any ram install).
BTW, I just noticed the neatest thing! The smiley sidebar of the advanced message editor here no longer causes safari to peg the computer! On the G5 dual core I had it would get up to 80% cpu usage easily.
|
|
MacBook Pro (Mid 2007), 2.4Ghz, 2GB DDR2-667Mhz, 160GB, Superdrive, Nvidia Geforce 8600M GT w/256MB, 15.4" WXGA+ LCD
|
| |
|
|
|
 |
|
 |
|
Addicted to MacNN
Join Date: Jun 1999
Location: Las Vegas, NV, USA
Status:
Offline
|
|
RAM speed is a measure of how fast the memory can guarantee the availability of the data after it has been given an address. That's why it is OK to use faster RAM than specified--because the data will be there when the CPU is ready to read it. Using slower RAM means that the correct data may not be there when the CPU reads the data out lines from the memory.
As far as I know, the memory bus does not "run" slower with slow RAM. The speed of the bus is function of the clock only. The risk is that CPU will get bad data.
Now many memory modules may actually be able to put the correct data out faster than they have to. But they are not guaranteed to work. This is somewhat analogous to the fact that some processors can safely run faster than they are spec'd to.
I would not take the risk, personally. Use the correct RAM.
Chris
|
|
|
| |
|
|
|
 |
|
 |
|
Posting Junkie
Join Date: Oct 2005
Location: Houston, TX
Status:
Offline
|
|
Originally Posted by chabig
RAM speed is a measure of how fast the memory can guarantee the availability of the data after it has been given an address. That's why it is OK to use faster RAM than specified--because the data will be there when the CPU is ready to read it. Using slower RAM means that the correct data may not be there when the CPU reads the data out lines from the memory.
As far as I know, the memory bus does not "run" slower with slow RAM. The speed of the bus is function of the clock only. The risk is that CPU will get bad data.
Now many memory modules may actually be able to put the correct data out faster than they have to. But they are not guaranteed to work. This is somewhat analogous to the fact that some processors can safely run faster than they are spec'd to.
I disagree. RAM is rated for a given speed, the maximum it is designed to operate at. Chipsets are also designed to support certain speeds. The actual speed it operates at is the lower of the maximum speed that the RAM supports and the maximum speed that your chipset supports.
|
|
|
| |
|
|
|
 |
|
 |
|
Addicted to MacNN
Join Date: Jun 1999
Location: Las Vegas, NV, USA
Status:
Offline
|
|
Originally Posted by mduell
I disagree. RAM is rated for a given speed, the maximum it is designed to operate at. Chipsets are also designed to support certain speeds. The actual speed it operates at is the lower of the maximum speed that the RAM supports and the maximum speed that your chipset supports.
You may be right. I'm a EE and what I stated was the way things worked 20 years ago. Things have certainly changed in those intervening years. But to work the way you suggest, the RAM module and the motherboard chipset would have to communicate using a protocol and negotiate a speed, then the chipset would have to adjust the clock accordingly. I somehow doubt that this level of intelligence is built into such a low-level system, but it's entirely possible.
Chris
|
|
|
| |
|
|
|
 |
|
 |
|
Administrator 
Join Date: Apr 2001
Location: San Antonio TX USA
Status:
Offline
|
|
Frequently, RAM controllers (this is generic-I have no data on the Intel iMac) will query the installed RAM for its rated parameters and configure themselves to run at the highest guaranteed speed that fits all installed DIMMs. So if you have a mix of RAM with different rated speeds, the RAM controller may, for stability's sake, throtle the RAM clock down to match the lowest rated speed. This is hardware and firmware dependent, but it appears to be both very common and very effective.
On many "gaming-focused" PC motherboards, you can tell the board to run RAM at whatever speed you want, just as you can overclock the processor. The tradeoff is often problems with stability. Macs just don't do stability tradeoffs, so I am assuming that they play it safe with the RAM clock.
|
|
Glenn -----
OTR/L, MOT, Tx
|
| |
|
|
|
 |
|
 |
|
Posting Junkie
Join Date: Oct 2005
Location: Houston, TX
Status:
Offline
|
|
Originally Posted by chabig
You may be right. I'm a EE and what I stated was the way things worked 20 years ago. Things have certainly changed in those intervening years. But to work the way you suggest, the RAM module and the motherboard chipset would have to communicate using a protocol and negotiate a speed, then the chipset would have to adjust the clock accordingly. I somehow doubt that this level of intelligence is built into such a low-level system, but it's entirely possible.
I don't believe it's a negotiation so much as the chipset can ask the RAM what clockrates and timings it supports (e.g. 3-3-3-8 at 200Mhz and 2.5-3-3-7 at 166Mhz and 2-2-2-6 at 133Mhz), and then the chipset picks the best available (or lets the user specify).
edit: Erm, yea, I should have read ghporter's post first; I'm saying the same thing he is.
|
|
|
| |
|
|
|
 |
 |
|
 |
|
|
|
|
|

|
|
 |
Forum Rules
|
 |
 |
|
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
|
HTML code is Off
|
|
|
|
|
|
 |
 |
 |
 |
|
 |
|