how-to block ads
|
mfrosty
Premium
join:2003-11-05
Crofton, MD
 ·Comcast
edit:
March 8th, @10:03AM
| Ultimate RAM guide
Ok, let me first start by saying I cannot take credit for writing this guide. It was written my a guy who goes by the name D-Rock on Extreme Overclocking forums. It's a great guide for both beginners and seasoned overclockers. I was so impressed with it and it contains SO much useful information that I wanted to post it here. (Trying to bring back the once great overclocking forum here) Anyway here it is, enjoy. Much props to D-Rock for putting in the time to do this.
Okay, where to begin, lets start by talking about the differences between DDR, DDR2, and DDR3.
DDR: DDR stands for double data rate. It was one of the first major breakthroughs in RAM technology that allowed for RAM to send and recieve data two times per clock cycle.With data being transferred 64 bits at a time, DDR SDRAM gives a transfer rate of (memory bus clock rate) × 2 (for dual rate) × 64 (number of bits transferred) / 8 (number of bits/byte). Thus with a bus frequency of 100 MHz, DDR-SDRAM gives a maximum transfer rate of 1600 MB/s. DDR DIMMS have 184 pins as apposed to the 168 pins on SDRAM DIMMS and 240 pins on DDR2 and DDR3 DIMMS.
DDR2: Since DDR stands for double data rate, by conjecture im sure you know that DDR2 stands for double data rate two. Like DDR before it, DDR2 cells transfer data both on the rising and falling edge of the clock (a technique called "dual pumping"). The key difference between DDR and DDR2 is that in DDR2 the bus is clocked at twice the speed of the memory cells, so four words of data can be transferred per memory cell cycle. Thus, without speeding up the memory cells themselves, DDR2 can effectively operate at twice the bus speed of DDR. This however comes at a cost, the latencies of these chips are greater, usually in the 4 to 6 range. As a side note it important to recognize that DDR2 DIMMs are not designed to be backwards compatible with DDR DIMMs. The notch on DDR2 DIMMs is in a different position than DDR DIMMs, and the pin density is slightly higher than DDR DIMMs in desktops. Another advantage that DDR2 brings to the table is that it is more energy friendly and therefore a more economical and efficient technological development. This is due primarily to die shrinkage.
DDR3: Once again by conjecture I'm sure that you have realized that DDR3 stands for double data rate three. Its primary benefit is the ability to run its I/O bus at four times the speed of the memory cells it contains, thus enabling faster bus speeds and higher peak throughputs than earlier technologies. This once again comes at a cost of latency. Also, the DDR3 standard allows for chip capacities of 512 mebibit to 8 gibibit, effectively enabling memory modules of maximum 16 gibibyte in size. DDR3 memory comes with a promise of a power consumption reduction of 30% compared to current commercial DDR2 modules due to DDR3's 1.5 V supply voltage, compared to DDR2's 1.8 V or DDR's 2.5 V. The main benefit of DDR3 comes from the higher bandwidth made possible by DDR3's 8 bit deep prefetch buffer, whereas DDR2's is 4 bits, and DDR's is 2 bits deep. For almost all intents and purposes the DDR3 technological advancement was not nearly as dramatic and ground breaking as the step from DDR to DDR2. Once again, DDR3 is not designed to be backwards compatible with DDR2 because DDR3 DIMMs have 240 pins which happens to be the the same number as DDR2, and are the same size, but are electrically incompatible and have a different key notch location. Some other new things that the DDR3 technology implements are the introduction of asynchronous RESET pin, and On-DIMM Mirror friendly DRAM pin outs.
Advantages compared to DDR2
Higher bandwidth performance increase (up to effective 1600 MHz)
Performance increase at low power (longer battery life in laptops)
Enhanced low power features
Improved thermal design (cooler)
Disadvantages compared to DDR2
Commonly higher CAS Latency
Generally costs much more than equivalent DDR2 memory.
How data is transferred from memory to CPU :
1) The CPU sends a signal specifying the memory row and bank that it wants to access via the RAS line.
2) After a period of time (RCD) the CPU sends a signal to the CAS line, specifying the column to be accessed.
3) After CAC time period, the data moves to the output line.
4) The CPU expects the data to 'appear' upon a specific clock cycle after sending the request.
__________________
************************************************** **************************************************
Next on the list will to be to explain the all important RAM timings. RAM timings have the potential to enhance performance of RAM by two or three times but on the contrast scenario have the ability to leave your system in a completely non-functioning state. Read all of this section before you go tweaking away. As a side note that should be mentioned, when you see advertised timings, the 4 that are shown are the ones that hold the only viable performance gains. The other timings could possibly improve memory performance or reliability but these are not very plausible means of improving RAM performance. There is little to no really helpful documentation on those timings anyway.
CAS: CAS stands for column address strobe latency referring to the column of the physical memory location in an array (comprised of columns and rows) of capacitors. CAS Latency (CL) is the time (in number of clock cycles) that elapses between the memory controller telling the memory module to access a particular column in the current row, and the data from that column being read from the module's output pins. Most people think that this is the most influential timing that affects RAM performance. This assumption is wrong but not very far off. CAS has a large impact on the performance of RAM but it has been proven that tRP and tRCD produce more of a substantial gain. e.g.: 2.5-3-3-8 The bold “2.5” is the CAS timing.
Normal setting per technological implementation:
DDR: common settings are 2, 2.5, and 3.
DDR2: common settings are 4, 5, and 6.
DDR3: common settings are 7,8, and 9. (5&6 plausible)
tRCD: tRCD stands for RAS to CAS delay (Row Address Strobe to Column Address Strobe). This is one of the more influential timings. This is the amount of time in cycles for issuing an active command and/or the read/write command. e.g.: 2.5-3-3-8 The bold “3” is the tRCD timing.
Normal setting per technological implementation:
DDR: common settings are 2, 3, and 4.
DDR2: common settings are 3, 4, and 5. (6 plausible)
DDR3: common settings are 5, 6, 7, 8, and 9.
tRP: tRP stands for row precharge (delay). This again is one of the more influential timings. This basically is the minimum amount of time between active commands and the read/write of the next bank on the memory module. e.g.: 2.5-3-3-8 The bold “3” is the tRP timing.
Normal setting per technological implementation:
DDR: common settings are 2, 3, and 4.
DDR2: common settings are 3, 4, and 5. (6 plausible)
DDR3: common settings are 5, 6, 7, 8, and 9.
Command Rate: Also called CPC (Command Per Clock). The amount of time in cycles when the chip select is executed and the commands can be issued. The lower (1T) the faster the performance, but 2T is used to maintain system stability. On Intel based machines, 1T is always used where the number of banks per channel are limited to 4. Most intel based motherboards don't cooperate well with 1t command rates even if the memory module is able to support it. Most times it will produce memory corruptions (especially with overclocking) which results in very poor performance.
(the following timings in my opinion have yeilded little to no performance gains as to the default (auto) setting.)
tRC Timing: Row Cycle Time. The minimum time in cycles it takes a row to complete a full cycle. This can be determined by; tRC = tRAS + tRP. If this is set too short it can cause corruption of data and if it is to high, it will cause a loss in performance, but increase stability.
tRRD Timing: Row to Row Delay or RAS to RAS Delay. The amount of cycles that it takes to activate the next bank of memory. It is the opposite of tRAS. The lower the timing, the better the performance, but it can cause instability.
tRFC Timing: Row Refresh Cycle Timing. This determines the amount of cycles to refresh a row on a memory bank. If this is set too short it can cause corruption of data and if it is too high, it will cause a loss in performance, but increased stability.
tRW Timing: Write Recovery Time. The amount of cycles that are required after a valid write operation and precharge. This is to insure that data is written properly.
tRTW/tRWT Timing: Read to Write Delay. When a write command is received, this is the amount of cycles for the command to be executed.
tWTR Timing: Write to Read Delay. The amount of cycles required between a valid write command and the next read command. Lower is better performance, but can cause instability.
tREF Timing: The amount of time it takes before a charge is refreshed so it does not lose its charge and corrupt. Measured in micro-seconds (µsec).
tWCL Timing: Write CAS number. Write to whatever bank is open to be written too. Operates at a rate of 1T, but can be set to others. It does not seem to work with other settings than 1T on DDR. DDR2 is different though.
Timings vs. Speed (throughput):
Now the question has been raised, which yeilds better performance, tighter timings or higher speed? Well looking at it from a theoretical and mathematical standpoint the answer would be like this. If you used your RAM very lightly and did not use RAM heavy apps then tighter timings would yield better performance. This is because the time to access RAM has been decreased. If you do use RAM hungry apps then timings would be over-ruled by overall throughput. From a realistic standpoint, most all operating systems use enough RAM to justify that higher throughput would result in better performance than tighter timings. Of course the most general consensus would be to have a nice combination of both. If you want to see just how much RAM you are using in real-time then there is a cool little freeware app that will do just that. Its called meminfo and there is no need to install it. Coincidently it barely uses any RAM. As a side note I think it's worth mentioning that timings have relatively little effect on Intel based systems anyway, the benefit of low timings is really that of AMD based systems. The reason for this? Well I couldn't tell you, I really don't know why that is but AMD systems see a really noticeable change with CAS latency.
************************************************** **************************************************
So the next section I will discuss is overclocking. I will discuss the procedure for overclocking as well as precautions and numerical and real life differences. Ram overclocking is very easy and is very ridden of guess work. Most of these figures will be able to be found using a simple household calculator. Most of the overclocking process (for RAM at least) is repetitive trial and error with memory testing in between adjustments. If that doesn't sound like your particular brand of whiskey then feel free to skip over this section. This is non-essential information. Note that in this section it is assumed that you have non-OEM parts in your PC and that you are not worried about voiding warranties.
For this section I am only going to discuss the conventional method of overclocking the ram, which involves the BIOS and will only work for motherboards that support manual settings of either the FSB or the RAM frequency or both. Also I should mention that this is my knowledge based on Intel based boards only because I have absolutely no experience with anything that is AMD.
Okay, it basically works like this: The first step in overclocking the RAM is to find out what speed your FSB is running at. We are going to go with 800mhz as an example throughout this tutorial.
When you have a FSB of 800mhz, and your RAM is rated for 800mhz then your ram would be running at a ratio of 1:1 of CPU:RAM. If you were to raise the front side bus to 850mhz then not only would you overclock your ram you woud also overclock your CPU (your ram would effectively be running at 850mhz) This however IS NOT TRUE FOR THE CPU! The front side bus of the cpu is not how fast your CPU is running. This is true because the CPU uses multipliers.
CPU speed is calculated like this:
FSB/4*(cpu multiplier)= Speed in mhz
800/4*(9 ex.)= 1800mhz or 1.8ghz
Here are come charts that have various multipliers and dividers:
»www.overclock.net/amd-memory/146···ers.html
Some motherboards allow you to set different ratios of CPU:RAM which allow you to overclock the CPU and RAM at propotionally different intervals. If you had a FSB of 533mhz and you wanted your RAM to run at 1066mhz all you would have to do is set a ratio of 1:2 (CPU:RAM) Playing with these ratios will give you different overclocks. 2:1 CPU:RAM synchronicity offers best performance on Intel systems the only exception being that loosening timings to achieve higher FSB will offer better performance than tight timings with limited FSB. When overclocking, highest CPU FSB is the ultimate goal. As you will read later in this post though, RAM running at a higher FSB than the CPU yeilds absolutely no performance gains. You must have the FSB running at a 2:1 ratio for the most optimal settings. This is because it is dual data rate, it operates at twice the speed of the cpu clock so if the ram were at 400mhz that would be optimal for a cpu with a fsb of 800mhz.
The procedure is as follows: This is where the redundancy comes rollin on in. First leave your timings as loose as possible and find the max mhz that your RAM will run at, then tighten the timings. To do that, raise the FSB buy say 10 or 15mhz each time, run memtest86+ all the way through, and repeat. When you start to see errors you have two options; you can decrease your overclock until you stop recieving errors or you can increase the voltage to the memory and keep going up until you reach the same roadblock. The process is the same for the CPU, except that the CPU frequency increases at a different rate. (which is FSB/4*(CPU Multiplier)) It will most likely be necessary for you to use a CPU:RAM ratio or to raise the voltage of the CPU when you overclock the RAM, assuming that you want to achieve relatively high RAM overclocks. ***NOTE THAT YOU MAY MAX OUT YOUR MEMORY CONTROLLER WHICH WILL HAULT YOUR OVERCLOCKING AT A CERTAIN SPEED***
***Running RAM at a slower rate than your CPU hinders performance, running RAM at a higher rate than CPU pretty much has no effect a 1:1 ratio is optimal***
***EDIT***Theoretically, if you get your RAM to run at the same effective bus speed as the CPU bus clock then that is optimal because anything else would just be overhead. However, that theory was just proven wrong because I have just been shown numerical evidence that supports the theory that memory performance still improves even after running faster than the CPU. The difference is so minute however that it would yeil no real world performance differences. It is only for numerical sake.
It may also be necessary to increase more voltages than just RAM (vdimm), when increasing the FSB you might need more power after a substantial increase. Some higher end boards allow voltage adjustments for the FSB as well as the north and south bridge. Without enough voltage to these things this will also cause instability so it would be a good idea to try these when RAM voltage increases aren't working any more.
RAM as it relates to AMD's K8 and others:
The integrated memory controller (IMC) on the K8 is definitely one of its greatest features. Since the controller is literally a part of the CPU die, it operates at the exact same frequency as the processor core. This very high speed operation coupled with direct access to the main memory contributes to the low memory latencies experienced on the K8 platforms. Obviously standard DDR and DDR2 memory can not operate at a 1:1 ratio to the memory controller. If that were possible, it would require 5,600MHz DDR for AMD's top models. In order to keep the memory operating at a more reasonable frequency, the memory operates at a set fraction of the CPU's frequency. The memory controller uses a 'whole number' divisor to achieve this. Take a 939 processor for example—the FX60. Operating at 2600MHz, its default memory frequency is 200MHz (PC3200 or DDR400). To obtain a 200MHz memory frequency, the '13' divisor or 'CPU/13' is used. This works very cleanly for all PC3200 based A64s on both the 754 and 939 platforms. Below are a few examples.
Default memory division for socket 754/939 processors:
CPU Frequency Memory Divisor (For DDR400) Resulting Memory
2800MHz CPU/14 200MHz
2600MHz CPU/13 200MHz
2400MHz CPU/12 200MHz
2200MHz CPU/11 200MHz
2000MHz CPU/10 200MHz
1800MHz CPU/9 200MHz
Since the standard memory frequency for AM2 processors is now 400MHz or 333MHz (DDR2-800 and DDR2-667), the memory controller is still forced to use a whole number divisor of the CPU frequency to obtain these frequencies.
Unfortunately, there are not always whole number divisors that can give you the expected 400MHz or 333MHz. Take the 5000+ for example. If you plug some numbers in your calculator, you'll notice that CPU/6.5 would be required to obtain a clean 400MHz memory frequency. Since this is not possible, the memory controller will automatically round the divisor up to a value of 7. So 2600/7 gives us a memory clock speed of about 371MHz. To make matters even more confusing, there are several AM2 models that can run their memory at a clean 400MHz, like the 4800+. The 'CPU/6' divisor makes it possible for the 4800+ to keep it's DDR2 running at the proper 400MHz frequency. Depending on the model you choose, your memory frequency can vary. Below is a table that outlines the actual memory frequencies you can expect:
Default memory division for socket AM2 processors:
CPU Frequency Memory Divisor (For DDR2-800) Resulting Memory
2800MHz CPU/7 400MHz
2600MHz CPU/7 371MHz
2400MHz CPU/6 400MHz
2200MHz CPU/6 367MHz
2000MHz CPU/5 400MHz
CPU Frequency Memory Divisor (For DDR2-667) Resulting Memory
2800MHz CPU/9 311MHz
2600MHz CPU/8 325MHz
2400MHz CPU/8 300MHz
2200MHz CPU/7 314MHz
2000MHz CPU/6 333MHz
1800MHz CPU/6 300MHz
So how far will my RAM overclock? Well that all depends on three things primarily. The first is your memory controller, you have to have a memory controller that will support higher DRAM frequencies then what is currently being run on the PC. Next is the IC that the memory module uses. The best (at the current time) by general consensus seem to bee the Micron D9 series of chips(DDR2). Most of these chips have the ability to reach over 1000mhz easily. The third most important thing that will affect overclocking is the PCB, (printed circuit board). This is the most overlooked aspect of RAM. Tests have shown that high end PCB's will yeild as much as 90mhz more performance if not more! Brainpower pcb's and 6 layer pcb's are the best. If you want to know what IC or PCB that your ram is using or if you want to know what to buy for overclocking I highly recommend that you visit the sites below for a reference. IC's highlighted in green are good for overclocking, IC's highlighted in red are bad for overclocking, and non-highlighted IC's are moderate.
Reference for
DDR: »ramlist.ath.cx/ddr/
DDR2: »ramlist.ath.cx/ddr2/
DDR3: »ramlist.ath.cx/ddr3/
Well what about my RAM getting to hot???
Memory cooling has become very popular, most notably on video cards. The effectiveness of memory cooling on both system ram and video cards, however, is often cause for debate amongst forum posters. Does system memory get hot enough to require cooling? Depends on what you consider is hot. My opinion is that memory modules never build up enough heat, to require cooling. Even when overclocking, they still stay pretty cool. If extra cooling, puts your mind at ease, then go for it, but you can't necessarily expect better overclocking results or even any extensions in the life of your overclocked memory. Premier manufacturers such as Corsair, Mushkin, and OCZ ship their modules with heatspreaders across the chips. They look very nice and are often solid copper or aluminum. There are even RAM modules that sell with heatspreaders that carry liquid flow across them and are meant to be integrated into an already functioning water cooling system. Several other companies sell ram cooling kits, and other solutions for modules that come without cooling. Ram sinks are pretty much the same as standard heatsinks for graphics chips and CPUs, except they're a lot smaller and tailored for RAM chip sizes. Tests show these heatspreaders & kits to do VERY little as far as cooling the memory goes. With no real benefit, placing these cooling kits on memory modules is more for looks than for cooling, and I can appreciate that.
How much voltage is too much?
To put it simply, there is no definitive answer to this question, it varies in many aspects, like IC, PCB, vendor, rating, and RAM type. As a general rule of thumb though it is not good to expose your RAM to voltages that are increased 2+ volts past stock for long periods of time.
Other things to know about RAM overclocking.
You should know that not only does RAM overclocking void most every warranty on the planet, it also will shorten the life span of your RAM. The two main factors that influence this are heat and voltage. Higher voltages usually do more damage then heat but both should be considered equally important.
KNOW THIS!!!
Just because Johnny across the street has his RAM overclocked to 5 million mhz that doesn't mean that yours will do that too just because you have the same RAM. RAM overclocking and its overall outcome have more factors than just the RAM itself.
************************************************** **************************************************
In this section I will cover some very common questions about RAM and some if its workings. This section will contain a generalized FAQ type style as well as certain terminologies explained.
Dual Channel, what exactly is dual channel? Take note that the memory isn't dual channel, the platform is. In fact there is no such thing as dual channel memory. Rather, it is a memory interface composed of two (or more) normal memory modules coordinated by the chipset on the motherboard. When referring to dual channel memory, it is referring to how the motherboard uses RAM. Basically, If you have two RAM channels on your motherboard (dual channel) and 512mb of RAM in each channel then each stick can hold completely different information and perform different tasks at the same time. This allows for the system to use large amounts of memory alot more effectively. Think of it as a 2 x 2-lane-highways, each highway represents a "channel". Cars can travel to and from a destination while the other highway has other cars going to and from a destination. This is a general idea of how dual channel works. If your mobo supports dual channel memory, then take advantage of it. Dual channel memory does not necessarily need to be purchased in a dual channel "kit", but you are best to install two identical matching sticks when running dual channel. If running dual channel with two speeds of RAM, your mobo will only operate in dual channel mode at the speed of the slowest stick. Dual channel memory greatly increases memory throughput (ideally it is supposed to double memory bandwidth, but in reality, this is not true).
What are the requirements for dual-channel? The requirements for dual channel are pretty simple. You need a motherboard that has dual-channel capability and you **usually** need two identical RAM modules. Some motherboards are more lenient than others. Some let you have two chips that have the same density and speeds and some require you to have everything matching right down to the serial number, these are called "factory matched pairs".
Well who made my ram chips or IC's?
Well that can usually be answered one of two ways, one of which is alot more certain than the other. The most definitive way to find out what IC your memory module uses is to just look at the module itself. Take the RAM out of your computer and look at the number-code that is etched into the black chips, then search that number on google. If you have a heat-spreader covering the black chips you can either take it off (which usually voids the warranty) or you can move on to the less definitive method. The less definitive method is to use an online chart that has most RAM IC's documented but is missing a few and needs some updating. They can be found here for:
DDR: »ramlist.ath.cx/ddr/
DDR2: »ramlist.ath.cx/ddr2/
DDR3: »ramlist.ath.cx/ddr3/
Well what IC's are good and what IC's are bad?
That really depends on what you want to do with the memory module, all IC's are generally fine for common use and are manufactured so they will work just fine with the advertised settings. If you however are not going to be using them in a general manor and tend to overclock quite frequently then there are major differences in the IC's. For the DDR series of RAM, the most notorious overclocker's IC is the Samsung TCCD. For the DDR2 series of chips, there are a few very reputable Micron IC's and they are all members of the Micron D9 series of IC's. For the DDR3 series one is really just as good as the other, there is not really much need to overclock these chips anyway because they are really fast enough out of the box. For reference here is a list of all the current Micron (common) IC's and their specs. Remember green indicates a good overclocker. Also you can find a picture based reference of common DDR2 chips here. Just in case you don't feel like leaving this post I will give you a basic run-down of some common DDR and DDR2 IC's.
DDR
Winbond BH:
(old + new school)
basically the ram that has turned into a myth of sorts. normally does around 240-270MHz.. not great in itself, but the thing to remember is that it can do it at 2-2-2 timings. the downside is that it typically needs 3.2-3.5v to do that. Also, higher latencies do not help much.. forget about cas3, it shouldn't boot. cas2.5 often isn't as stable, or simply does not get many more MHz. Try not to go higher than 3.5-3.6v unless benching, and be sure that your system is up to taking that much voltage for long term use.
UTT-CH:
(new ch)
known to be in OCZ VX, twinmosSP and some Mushkin Redline modules. It's similar to bh-5, in the fact that it likes lots of voltage and generally likes 2-2-2 timings. CH seems to scale with voltage better than BH, but it needs around 3.1v to to get down to 2-2-2 to begin with. Being a smaller process size, it's not as safe to pump CH dies as much voltage as BH. I wouldn't go over 3.4v for 24/7
Samsung TCCD/TCC5:
Been around for a little while. It had.. a cult like following for a while, which I find amusing, though it's for a good reason.. anywho, the premise of this stuff is that you can do really high speeds (often 300MHz+ with good sticks) and doesn't need much voltage. The downside is that it needs higher timings to get there. Expect to have to raise timings to 2.5-4-4 or 2.5-4-3 to get the most out TCCD/TCC5. Also be prepared to do a little bit of testing to find which voltage works best, some sticks actually start decreasing performance if you give them more than 2.7-2.8v, while some can take up to 3v. with newer tccd/tcc5, it seems that more than 2.8v will hurt then after prolonged periods of time, while the older chips can take a bit more, and actually scale really well at cas2 with lots of voltage, but die really fast if you do that.
Samsung UCCC:
Though it is kinda "value-oriented" as far as 64mb ICs go, proper binning can get you 250MHz 3-4-4-8 or 3-4-3-8. Typically top out around 270-280MHz, though sticks have been recorded at 300MHz. OCZ and G.Skill have a wide variety of sticks with these chips. Other manufacturers known to use them are Corsair, Mushkin, Teamgroup, and Crucial.
Micron 5b G:
this is sorta a mix between BH-5 and TCCD. With properly binned sticks, it can do 2-2-2 timings up to a moderate speed (typically 220-225MHz without going nuts with voltage). The sweet spot is usually with 2.5-2-2; 250MHz at 2.7-2.9v and 2.5-2-2 seems typical with 5b G, with more possible with additional voltage. It also scales well with timings too, though not as well as TCCD does, with 3-3-3 being best for hitting highest MHz (around 275-310 usually). 5b G can take 3.0-3.2v pretty reliably when properly cooled, though it will cut into it's lifespan a bit. if you get it in value ram, simply subtract ~10-40MHz from the above numbers, depending on your luck
Micron 5b D & 5b F:
Similar to 5b G, but is typically a 64mb IC instead of 32mb. Being a smaller process, 110nm for D and 95nm for F, don't give these as much voltage, around 2.9v and 2.7-2.8v seems to be best, and at 3-3-3. Be exceptionally careful with 5b D, as it seems to have a very high failure rate.
Infineon CE-5/CE-6:
These can be considered the BH series of 1GB sticks (though not to me), minus the high voltage. Though they are usually limited to around 260-270MHz, they make up for that with lower timings than the other 64mb IC's. Infineon chips are usually binned to 250MHz 3-3-2, which is not half-bad at all. Usually they will go up to 260-270MHz 3-3-2.
DDR2
Micron BT series
Most people refer to these as "fatbody", due to the large package size (92-ball IC). The label on the IC package is either D9DQW (266MHz at cas4) or D9DQT (333MHz at cas5). Likes relatively low timings, 3-2-2 to 4-3-2, and higher voltages. Will do moderate MHz, but getting over 500-550MHz seems somewhat rare. Long term use over 2.5v will most likely degrade stability.
Micron B6 series
These are BT's replacement. D9GKX is the 400MHz, cas5 binned part, D9GMH is 333MHZ, cas5 and D9GCT is 266MHz, cas4. On average, GKX hits the highest speeds and GCT does the lowest. Good all around chips, should do 350-450MHz at 3-3-3, 450-600MHz at 4-4-3 or 4-4-4 and >500MHz at 5-5-4 or 5-5-5. Scales well with voltage like BT, but being made on a smaller process, long term use over 2.3 to 2.4v is not recommended.
ProMOS
ProMOS's current generation of chips is a good value alternative, and shows up in some of the cheaper "performance" kits. My experience with it is that added voltage over 2.3-2.4v does next to nothing. Even so, going over 2.3v for 24/7 is not advised. Should do 300-400MHz at 3-3-3, 350-500MHz at 4-4-4, and may or may not gain more speed with timings looser than 4-4-4.
Will adding more RAM to my computer make it faster?
That depends almost entirely on how much memory the tasks you do on your computer take. If you are running simple or older programs that simply are not memory intensive, additional memory largely cannot help you. But over time operating systems and applications tend to require (and therefore benefit from) increasing amounts of memory, and in most cases it's hard to have too much. More memory will not increase the speed of the CPU, but it will reduce the time a CPU spends waiting for information from a hard drive. The operating system and applications will be able to load more of their data into ram at once, and the dependence on virtual memory (see virtual memory below) will be reduced. Since RAM provides data to a CPU faster than a hard drive, you will not have to wait as long for programs to execute in most cases. If you want your computer to run faster in nearly all cases, consider upgrading the CPU or overclocking. RAM is nothing more than an electronic storage area for data. RAM stores data like a hard drive but a hard drive is mechanical is consequently alot slower in data transfer than RAM is. So this makes data storage alot faster. When you have to write data to the hard drive it slows down your computer. You need enough RAM to meet your maximum usage, but no more than this. If a program requires more memory then you have installed, free hard drive space will be converted into virtual memory. So why not use your 500GB hard drive for memory? Simple - speed. Your system will crawl and games stutter if you have to use virtual memory.
Average RAM speed: 8000MB/s xfer rate, 4ns latency (nanoseconds as in one billionth of a second)
Average Hard Drive speed: 60MB/s xfer rate, 14ms access time (milliseconds, as in one thousandth of a second)
However, you do not want too much, because it puts additional strain on the memory controller, overclocks worse and costs more. You only need to satisfy your maximum usage.
What is virtual memory?
This is a method of extending the available physical memory (RAM) on a computer. It basically partitions off a section of a hard drive and uses it in place of RAM. It uses this section to read and write data that was too large to fit into RAM. Virtual memory is drastically slower than RAM. If you are having to use the swap file/page file at all then that is when more RAM would improve computer speed.
Burning in RAM, what is it and does it really work?
Burning in ram is basically like breaking in an engine. You control all of the operating variables and run it in that state for a long period of time. This allows for it to perform better once the operating variables are changed. Burning in RAM is said like this. You overclock your RAM, then you burn it in. All thats doing is letting the RAM get really really comfortable operating at those specific variables. This allows for it to be overclocked in the same proportion alot easier. Does it really work? Well the short answer is yes, it does. It yeilds some of the highest overclocks but this method is should only really be implemented if you are an "overclocking junkie" and you are looking for your true maximum overclock. Another more common use for it is when you hit a wall when overclocking. It is very redundant work though.
Can I use faster rated RAM than my motherboard supports?
Well of course you can, there is a drawback of course, your RAM will run as fast as your motherboard can handle and not at its full potential. In this particular equation, the motherboard is the limiting factor.
What programs can I use to check my memory?
There a few pretty reliable programs out there that will let you check different things about your memory. If you just want to know what type of memory you are using and the SPD specs of the modules then you can use CPU-z. If you want to benchmark your memory then Sandra Lite is a good choice, this is made by Sisoft. If you want to see if your memory is having errors then there are two options for this, there is the less reliable software option, and the more reliable bootable dos option. The software version is memtest. The more reliable and thorough version is memtest86+. It should also be noted that even though they are both under the name of memtest, the are not however made by the same organization.The reason that the software version is less reliable is because it doesn't have the ability to check the RAM that is being used by the operating system and currently running programs.
RAM defraggers, theory and actual impacts:
RAM defraggers in theory are designed to not "defrag" the memory but to push all of the RAM to a minimized and compressed level while in the RAM. This is actually a bad thing because it could lead to data corruptions and it is actually taking RAM away from the already running programs. I advise against these types of programs.
Whats the difference between buffered and unbuffered DIMMs?
High density DIMMs have lots of chips on them and therefore possess a higher capacitive load on the address and control signals in comparison to lower density DIMMs. Some designers use redrive buffers on the DIMM to boost the signals to reduce system loading when compared to the same high density module without buffers. But the buffers introduce a small delay into the electrical signal, so adding buffers to a standard density module would have the effect of slowing down the signal, compared to the same low density module without buffers. FBDIMMS are only common in server applications and are not necessary for desktop/workstation applications. The purpose of an FBDIMM is to preserve data integrity. (Maintaining data integrity in a server is essential.)
Whats the difference between 72bit and 64 bit/ what is ECC?
72 bit memory is commonly known as ECC memory. It has an additional 8 bits for Error Correction Check 64 bit memory is non-ECC. 72 bit configurations are typically found in 168 pin DIMMs whereas 64bit configurations are found abundantly.
Why can't i see/use 4gb of RAM?
Well this is because 32bit operating systems (most common type of os) only have enough memory address space for 4gb. This doesn't mean its all going to your RAM though. This also includes graphics RAM, networking peripherals, pci cards, and the like. You will only be able to see 3.2~3.5gb of RAM on a 32bit operating system. High end computer users might be using a 64bit operating system which natively supports addressing for much larger amounts of RAM. The current maximums for windows operating systems are as follows:
Windows XP Home: 4GB
Windows XP Professional: 4GB
Windows XP 64-bit: 128GB
Windows Vista Home Basic: 4GB
Windows Vista Home Basic 64-bit: 8GB
Windows Vista Home Premium: 4GB
Windows Vista Home Premium 64-bit: 16GB
Windows Vista Ultimate: 4GB
Windows Vista Ultimate 64-bit: 128GB+
Windows Vista 32-bit: 4GB
Windows Vista 64-bit: 128GB+
Additional Links and resources:
Wikipedia.org
DSLReports.com
RAMlist.ath.cx
techpowerup.com
Pcstats.com
extremeoverclocking.com
eclipseoc.com
cpuid.com
HCIdesign.com
sisoftware.net
memtest.org
buycomputermemory.com
crucial.com
overclock.net
icronic.com | |
Somnam
L33t.
Premium
join:2002-12-05
Mullica Hill, NJ
clubs: | excellent find. even stuff i learned from it | |
jsimmons
Premium,MVM
join:2000-04-24
Falls Church, VA | reply to mfrosty
Excellent overview. Well written and a solid reference. | |
Babar
Premium
join:2001-05-09
Washington | reply to mfrosty
Thanks, mfrosty - good stuff! got a link? | |
Spinnaker
Ok, But What Is The Speed Of Dark?
Premium,MVM
join:2002-12-08
Holden, ME
edit:
April 4th, @07:12PM
| reply to mfrosty
I'm not mfrosty  , but there's two links over at the EXTREME Overclocking Forums...
D-Rock posted one to the Intel Memory forum --»forums.extremeoverclocking.com/s···AM+guide and the other to the AMD Memory forum --»forums.extremeoverclocking.com/s···AM+guide
--
Sprint EVDO Rev. A, Novatel U720 Aircard, Kyocera KR1 router, Wired/Wireless LAN, 2-WinXP Pro w/SP2, 1-Win98SE | |
|
