ChristopherM
New member
Anyone try 'em? Opinions?
F
Fusion2
New member
ChristopherM said:
nope, sure haven't, there pretty, that's about it, cheesy memory reseller use them to hide crappy ram makes with "warranty void if removed" crosair, crucial, samsung etc do not come with sinks... it's all marketing imho...
ram gets warm, not hot, and that may factor into your systems overall temp, a side case vent with 80mm directed between the cpu and ram helps cool much more hardware than just the ram, i believe the o'l saying is, heat kills hardware, very true, hard drives are a good example, a case vent helps there also...
video card ram? that's another story...
Last edited:
M
Marquis
New member
The best RAM comes with them, but not for any good reason other than (supposed) aesthetics. They are useless.
F
Fusion2
New member
i stand corrected, i just remembered corsair's extreme series overclocking memory comes with the snap on heat sinks, in an overclocked state i'd be more concerned with cooling the north bridge and a few other small chipsets that get very hot before the ram, imho of course...
i ordered about 25 sticks of pc2700 - 3000 this spring and only the lowest rated ram cames with a heat sink, overclockers, life time warranty but, void if you remove the sinks, otherwise crucial 512 and corsair 512 came without sinks...
the ram cost right at $50 for 512mb when i purchaced it, very true, it is a commodity, and has a roller coaster price tag...
just thought i'd throw that tid bit in...
i ordered about 25 sticks of pc2700 - 3000 this spring and only the lowest rated ram cames with a heat sink, overclockers, life time warranty but, void if you remove the sinks, otherwise crucial 512 and corsair 512 came without sinks...
the ram cost right at $50 for 512mb when i purchaced it, very true, it is a commodity, and has a roller coaster price tag...
just thought i'd throw that tid bit in...
M
Marquis
New member
"i stand corrected, i just remembered corsair's extreme series overclocking memory comes with the snap on heat sinks, in an overclocked state i'd be more concerned with cooling the north bridge and a few other small chipsets that get very hot before the ram, imho of course..."
Perhaps your opinion should be a little less humble...
You're spot-on when it comes to overclocking. Just look at Abit's new IC7-MAX3 and it's active cooling of the power regulation circuitry. It's pure genius, though something that the hardcore have been doing since the days of the infamous Celeron 300A(perhaps before, but that was the first time I saw it).
The ramsinks on high-end memory is just a gimmick. It'd perform the same without them, but companies want to add shiny widgets to attract the uninformed (or at least to keep up with the corporate "Joneses" who are doing so).
Perhaps your opinion should be a little less humble...
You're spot-on when it comes to overclocking. Just look at Abit's new IC7-MAX3 and it's active cooling of the power regulation circuitry. It's pure genius, though something that the hardcore have been doing since the days of the infamous Celeron 300A(perhaps before, but that was the first time I saw it).
The ramsinks on high-end memory is just a gimmick. It'd perform the same without them, but companies want to add shiny widgets to attract the uninformed (or at least to keep up with the corporate "Joneses" who are doing so).
sweetnubs
New member
overclocking is just a gimmick and not worth the effort and time involved. Just kick out the extra cash and save your hardware from frying and software from corrupting. There are too many variables involved and any serious engineer worth his salt is more interested in stability.
C
charger
New member
I think they're a gimmick. I'm running a 512MB stick of Geil Golden Dragon RAM (no heatsinks, just a clear plastic cover). I've been running it at 440 MHz (it's rated for 400) and began to get a little concerned about it overheating. Turned off the computer, popped open the case, and touched it. Not even warm after 8 hours. Ditto the CPU heatsink and the northbridge heatsink. Overclocking may be a gimmick, but it sure makes my computer faster!
M
Marquis
New member
Overclocking is only a gimmick for those with unlimited budgets and/or those that don't understand the process.
The way CPUs are speed binned often has little to do with the speeds that they will operate at and more to do with the demands of the market. This is particularly true of mature lithography processes (such as Intel's and AMD's current .13 micron processes).
Besides, anybody that overclocks a CPU (or video card, etc.) and doesn't spend quite a bit of time testing for stability deserves whatever he/she gets (be it instability or fried components). I have about a dozen subsystem and whole system stress tests that each system I build goes through, overclocked or not. They all pass or the system gets fixed (or in the case of pushing the limits of O/Cing, reconfigured and retested).
I'm currently building a pair of Athlon XP systems for my parents (one as a DAW for my dad, one as a video editing workstation for my mom) that have Barton 2500+ processors running at 3200+ speeds. The difference in prices between the two is $232 from NewEgg. I don't consider that to be a small difference, especially when the CPUs are built in the same place using the same process and then "speed-binned".
To each their own, of course. I can also understand why those who make a living with their studios might be especially risk-averse.
The way CPUs are speed binned often has little to do with the speeds that they will operate at and more to do with the demands of the market. This is particularly true of mature lithography processes (such as Intel's and AMD's current .13 micron processes).
Besides, anybody that overclocks a CPU (or video card, etc.) and doesn't spend quite a bit of time testing for stability deserves whatever he/she gets (be it instability or fried components). I have about a dozen subsystem and whole system stress tests that each system I build goes through, overclocked or not. They all pass or the system gets fixed (or in the case of pushing the limits of O/Cing, reconfigured and retested).
I'm currently building a pair of Athlon XP systems for my parents (one as a DAW for my dad, one as a video editing workstation for my mom) that have Barton 2500+ processors running at 3200+ speeds. The difference in prices between the two is $232 from NewEgg. I don't consider that to be a small difference, especially when the CPUs are built in the same place using the same process and then "speed-binned".
To each their own, of course. I can also understand why those who make a living with their studios might be especially risk-averse.
F
Fusion2
New member
OC was cool in it's day, similar to racing a mhz tweak will at times help stablize a system that's not, tweak, as in 1 to 25mhz...
like i say, the comp waits on me much of the time anyway, OC is in it's twilight years imho...
next stop, 4000mhz
i'm glad to see users here are not marketing suckers...
i believe i'll stay at around 3ghtz for about 5 years before upgradeing, ahem, that is, if there's anything worth upgrading to at that time...
a mhz tweak will at times help stablize a system that's not, tweak, as in 1 to 25mhz...like i say, the comp waits on me much of the time anyway, OC is in it's twilight years imho...
next stop, 4000mhz
i'm glad to see users here are not marketing suckers...
i believe i'll stay at around 3ghtz for about 5 years before upgradeing, ahem, that is, if there's anything worth upgrading to at that time...
sweetnubs
New member
I'm quite aware of the process and familiar with "speedbinning" The problem is the amount of variables. Is the cpu at the proper temp, the gpu, the chipset, the hard drives, are you sending the right frequecies to your peripherals, hows the power supply doing now?, once you bluescreened what sofware is now corrupted, are you devloping hotspots on your motherboard, have you stress tested every component? Time to reset the cmos. Regardless if you do cool properly the life of your components are still reduced. To see any real benefits you need to invest in higher quality components, better heatsinks, bigger power supplies, faster fans which run quiet (you are recording right?), etc. This can cost as much money as builing a comparable non-overclocked rig. To do this properly takes vast amounts of time and research and still may result in failure, corrupted data, etc. which means even more time lost. Engineers simply do not have enough time for this. It is more of a hobby/sport and not a sensible solution for recording professional audio. However all the techniques used to stabalize a overclocked rig are very valuable for stablizing a non overclocked rig. I.e. monitoring your temperatures, stress testing, searching for hotspots, testing for proper airflow in your case, etc.
M
Marquis
New member
"like i say, the comp waits on me much of the time anyway, OC is in it's twilight years imho..."
Quite the contrary, I'm of the belief that there is a current golden age of overclocking similar to the Celeron 300A days. Anybody can pick up an Athlon XP Barton 2500+ or Intel P4 2.4C and run it at stock voltages and get quite the overclock (often with the stock heatsink, though I usually recommend a Zalman CNPS700Cu for noise purposes). Besides, there is usually a good time during any fab process for overclocking. Determining when that is takes experience, or some poking around the right Internet forums asking those who have it.
Maybe it's just that this is all so second-nature to me (I've been building systems for 20 years and overclocking for 19), but I never viewed it as a hassle. OK, so soldering in a new clock crystal was a bit of a hassle. But it was cool to have the fastest 8088 in the land.
Nowadays, though, a BIOS tweak or two is all it takes. Easy...
And sweetnubs, I would choose the same components for a mildly overclocked system (i.e., a 2500+ at 3200+ speeds or a 2.4C at 3GHz) that I would for a non-overclocked system anyway. I never buy marginal components. So there is no cost differential. Now, if I'm trying the uber-overclock so I can go brag on some OC forum, that's perhaps different.
As you well point out, stability testing should be done irrespective of your choice to overclock or not (at least for DIY systems, OEMs ostensibly do that kind of testing for you).
BTW, I'm not trying to evangelize here. As I said before, to each his/her own (especially when your livelihood is on the line). I just think that people should be informed what their options are, especially for us home recording hobbyists who don't have large budgets(not that comercial studios necessarily have lots of money to waste).
Quite the contrary, I'm of the belief that there is a current golden age of overclocking similar to the Celeron 300A days. Anybody can pick up an Athlon XP Barton 2500+ or Intel P4 2.4C and run it at stock voltages and get quite the overclock (often with the stock heatsink, though I usually recommend a Zalman CNPS700Cu for noise purposes). Besides, there is usually a good time during any fab process for overclocking. Determining when that is takes experience, or some poking around the right Internet forums asking those who have it.
Maybe it's just that this is all so second-nature to me (I've been building systems for 20 years and overclocking for 19), but I never viewed it as a hassle. OK, so soldering in a new clock crystal was a bit of a hassle. But it was cool to have the fastest 8088 in the land.
Nowadays, though, a BIOS tweak or two is all it takes. Easy...And sweetnubs, I would choose the same components for a mildly overclocked system (i.e., a 2500+ at 3200+ speeds or a 2.4C at 3GHz) that I would for a non-overclocked system anyway. I never buy marginal components. So there is no cost differential. Now, if I'm trying the uber-overclock so I can go brag on some OC forum, that's perhaps different.
As you well point out, stability testing should be done irrespective of your choice to overclock or not (at least for DIY systems, OEMs ostensibly do that kind of testing for you).
BTW, I'm not trying to evangelize here. As I said before, to each his/her own (especially when your livelihood is on the line). I just think that people should be informed what their options are, especially for us home recording hobbyists who don't have large budgets(not that comercial studios necessarily have lots of money to waste).
D
ds21
New member
I don't want this to get into a flame war, but I overclock, been doing it for about 8-9 years, saved some money, still have the first system and it works fine, it's comming back into vouge, once chip makers "overclock" their chips to the point where they have to look into new ways of manufacturing, then is when the diminishing returns come into play, but for certian chips it works flawlessly (if you know what your doing) I just do it because I can.
M
Marquis
New member
Hopefully nobody thinks this is or has ever been a flame war. I certainly don't view it as such.
However, I'm not sure I agree with your assessment with respect to when a good time to overclock is. Often the end of a process generation is the perfect time to do so, though it depends on what's coming over the horizon. It is at this point in the manufacturing process where almost all the chips coming off the line are basically the same. That means the low-end is just as good the high-end, it's just cheaper. Now is a perfect example with the 2.4C and 2500+. They are mature processes that still need to produce "lower" speed chips because most people won't buy $300+ processors.
Of course, I think we'll both agree that early generations of new processes are nearly universally the worst time to overclock.
"if you know what your doing"
Yes, good point. I often fail to mention this for some reason, though I am remiss to do so. Knowledge of any endeavour, of course, leads to greater levels of success.
However, I'm not sure I agree with your assessment with respect to when a good time to overclock is. Often the end of a process generation is the perfect time to do so, though it depends on what's coming over the horizon. It is at this point in the manufacturing process where almost all the chips coming off the line are basically the same. That means the low-end is just as good the high-end, it's just cheaper. Now is a perfect example with the 2.4C and 2500+. They are mature processes that still need to produce "lower" speed chips because most people won't buy $300+ processors.
Of course, I think we'll both agree that early generations of new processes are nearly universally the worst time to overclock.
"if you know what your doing"
Yes, good point. I often fail to mention this for some reason, though I am remiss to do so. Knowledge of any endeavour, of course, leads to greater levels of success.
D
ds21
New member
what I mean is the actual chips at the top of there limit, those chips that you can't really overlock that much because their already at their thermal limit.
F
Fusion2
New member
it's not very apperent but i was being slightly scarcastic with sweet, i've OC since the 300a and lost no hardware because of it, i remember fondly pushing my old duron 700 passed 1ghtz and the feeling of breaking that limit at the time before the public had years ago, i also managed my 300a to 450 when i received an abit av6 i believe it was but don't hold me on that as it was indeed, long ago as of lately i have a game box with xp1700/266 that i run at 167fsb and get some outragous clock over 500mhz or so, it's so common i don't even care what mhz it does run, as long as all systems specs are within a safe level and no corruption or lock ups occur... and of course, sisoft rock stable...
my audio systems are spec tweaked, not OC really, can't afford a lock up and data loss there...
the only real corruption i've ever seem was from pushing the memory timing to high, and just backing it down fixed that, your experences may have differed though...
yes you needed a big heat sink but that's where we are today anyway, the extra cost to OC was little if that...
in most cases you'll get a blue screen if you over do it or miss config the system attempting a given clock, now i'm not saying it's not possible to kill or corrupt something, it's just i've had a horse shoe up my rear end and oclocked heavily the last 4 or 5 years without loss and little issue, knock on wood
mhz have gotten so high the speed benfit is gone now imho, as the comp surely waits on me and not i on it, there's always the bench numbers to bicker over though...
peace...
Last edited:
F
Fusion2
New member
ds21 said:
cpu yes, ram no, there are large copper sinks or water cooling to deal with those issues...
true, they are towards end of the mhz limits as far as 32bit goes...
i never thought all these clockers would hang here at HR, heh...the web is surely, a big place...
Last edited: