RAM, 1866 is obviously better than 1600 isn't it?

[Nec_V20]

Interesting discussion. Nec_V20. I feel this discussion is 13 years too late. Why? In 2000, SDRAM was in virtually every computer you could buy, and the last time I ever remember cas latency and speed causing a noticeable difference in my computer's performance (noticeable to me, not just in benchmarks).

Heck, the laptop I am on right now, Lenovo Thinkpad W530 with 8GB DDR3, is running one stick of RAM, and is one of the faster machines I have ever used. If it had a faster hard drive, like in the desktop in my sig or an SSD, it would be the fastest (and my desktop does run dual channel).

I think you need to come to grips with the times here. And, as others have stated, making statements as you have done would require some proof, which you don't seem to have.

If you look at the price vs. performance of RAM compared to CPU or graphics cards, then the large difference in the high price for the relatively minor and for the most part unnoticeable performance gain just isn't worth it.

You won't notice any difference under most circumstances between 1600 and 2400 RAM except in your wallet.

Now if you have something to contribute to the discussion then please do so. Jumping on the ad hominem bus is just stupid, especially as the argument put forward in my original post has already been more than adequately documented on this thread, if you had just bothered yourself to look.

If your Lenovo W530 is so great then you should document that otherwise it is just a hollow boast on your part. After all you were just making an unfounded statement and why should I take you at your word?

One reason for me to take you at your word is that I don't automatically assume that you are posting here in bad faith.

 

[Ketchup]

Jumping on the ad hominem bus is just stupid, especially as the argument put forward in my original post has already been more than adequately documented on this thread, if you had just bothered yourself to look.

Nope, my argument is with your statements. Nothing against you personally. I read your OP, did you? Maybe your links got broken, or you consider your own thoughts/words your proof. The other posters in this thread have already provided the proof you couldn't, it just went against your argument, so I am sure it was promptly ignored.

I am not giving proof, because I am not trying to say something everyone else disagrees with. The ball is in your court, as it were.

 

[dbcooper1]

As I said before, the difference in price between 16GB (2*8GB) 1600 CL9 RAM and 16GB (2*8GB) 2400 CL10 RAM is the difference in price between the two Haswell processors i5-4670k and i7-4770k

Far more bang for your buck from Hyperthreading and the additional cache of the 4770k than from difference in RAM in most non-synthetics uses.

 

[Nec_V20]

Nope, my argument is with your statements. Nothing against you personally. I read your OP, did you? Maybe your links got broken, or you consider your own thoughts/words your proof. The other posters in this thread have already provided the proof you couldn't, it just went against your argument, so I am sure it was promptly ignored.

I am not giving proof, because I am not trying to say something everyone else disagrees with. The ball is in your court, as it were.

Actually no, I didn't read my original post, that is why it is riddled with spelling mistakes and grammatical errors. I was just too embarrassed and I hoped you wouldn't notice.

So what's your beef? I posted something that was in the realms of Captain Bloody Obvious. Those who know confirmed it.

If I told you I encountered daylight this afternoon I wouldn't expect you to post back to me to send you a photo to prove it.

Although it is obvious to people such as myself (i.e. people who have had years of experience), there are a lot of people who could get caught up in the marketdroid and salescritter hype (read lies) espousing the miraculous properties of "spectacularly fast" RAM drastically improving performance and be conned into spending money on a promise more properly realised by investing their money in some other piece of hardware (such as a more potent CPU - an i7-4770k instead of an i5-4670k - for instance) which actually would give a noticeable performance boost.

 

[Ketchup]

Actually no, I didn't read my original post, that is why it is riddled with spelling mistakes and grammatical errors. I was just too embarrassed and I hoped you wouldn't notice.

So what's your beef? I posted something that was in the realms of Captain Bloody Obvious. Those who know confirmed it.

If I told you I encountered daylight this afternoon I wouldn't expect you to post back to me to send you a photo to prove it.

Although it is obvious to people such as myself (i.e. people who have had years of experience), there are a lot of people who could get caught up in the marketdroid and salescritter hype (read lies) espousing the miraculous properties of "spectacularly fast" RAM drastically improving performance and be conned into spending money on a promise more properly realised by investing their money in some other piece of hardware (such as a more potent CPU - an i7-4770k instead of an i5-4670k - for instance) which actually would give a noticeable performance boost.

I think I see the problem here. There was a communication breakdown of what I meant when I typed the post, and how you took it. I'll take the blame for this. Since we are several posts down now, I am going to leave it. If you want me to try again, I will. If not that's fine too.

Do try to lighten up a little though. It's just a forum.

 

[Nec_V20]

I think I see the problem here. There was a communication breakdown of what I meant when I typed the post, and how you took it. I'll take the blame for this. Since we are several posts down now, I am going to leave it. If you want me to try again, I will. If not that's fine too.

Do try to lighten up a little though. It's just a forum.

Looking back I feel I should apologise.

We appear to have been talking at cross purposes which is as much my fault as anyone else's.

 

[Ketchup]

Looking back I feel I should apologise.

We appear to have been talking at cross purposes which is as much my fault as anyone else's.

Appreciate the PM. We're cool.

 

[Blain]

I've been building systems for many years. I've been through the "phases". What I've learned is that at the end of the day I want quality for the dollars I spend.
It seems like many people are so fixated on the measurable or perceived performance benefit (or lack thereof), that they fail to step back and look at the big picture...

Is this good quality memory that will last me for the long haul, or is it a flash in the pan that will fail prematurely?

 

[Nec_V20]

I've been building systems for many years. I've been through the "phases". What I've learned is that at the end of the day I want quality for the dollars I spend.
It seems like many people are so fixated on the measurable or perceived performance benefit (or lack thereof), that they fail to step back and look at the big picture...

Is this good quality memory that will last me for the long haul, or is it a flash in the pan that will fail prematurely?

My main system has been up an running now 24/7 for 1,005 days 6 hours, so I hear you.

Every four or five years I build up a new main system for myself. I agonise somewhat over the components but when it is complete it becomes the "Grizwald" (my current system is "Grizwald 7", and the NAS I recently built is called "Halfpint"). They become family.

The longest lasting Grizwald was Grizwald 2, which was based on a Harris 25MHz 286 and that lasted me through the 386 era (although only clocked at 25MHz it ran the OS and software of the day faster than any 386). The next Grizwald (Grizwald 3) I built was based on the 486 DX50.

After that came Grizwald 4 which was my first dual CPU (Pentium MMX 200MHz) on a Tyan Tomcat board.

I am still deciding whether the system I will be plugging in and starting up today which is based on an i7-4770k and a Gigabyte Z87X-UD5H board will become Grizwald 8.

As for RAM, I have had very good experiences with Patriot Memory 1600MHz Viper 3 Black Mamba CL9 series. Both with the 4GB and the 8GB modules.They just seem to feel comfortable in every system I have placed them in.

Like you, I build to last and not to brag.

 

[Makaveli]

Interesting thread.

I have a 1600 Mhz Cas 7 kit if you want any benchmark numbers to compare.

 

[Ketchup]

My main system has been up an running now 24/7 for 1,005 days 6 hours, so I hear you.

What OS is on that? 7 is the longest OS I have kept running between reboots, but I usually reboot it every week or so, and during the summer it sleeps/hibernates a decent amount.

Also, over the years I personally have had only 1 instance of a RAM stick going bad. I have fixed that issue for other people, but only once personally. And that computer is a Compaq I got for my parents 7 years ago (I still use it as my media/data server) so I'd say that's pretty dang good. Don't feel bad for Mom and Dad though, lol. A couple years ago when we came over to visit I surprised them with a Core 2 Quad rig I built with spare parts. Mom was quite pleased.

 

[Nec_V20]

What OS is on that? 7 is the longest OS I have kept running between reboots, but I usually reboot it every week or so, and during the summer it sleeps/hibernates a decent amount.

Also, over the years I personally have had only 1 instance of a RAM stick going bad. I have fixed that issue for other people, but only once personally. And that computer is a Compaq I got for my parents 7 years ago (I still use it as my media/data server) so I'd say that's pretty dang good. Don't feel bad for Mom and Dad though, lol. A couple years ago when we came over to visit I surprised them with a Core 2 Quad rig I built with spare parts. Mom was quite pleased.

Sorry, of course I do a reboot every week or so, if updates don't require one. It just gets rid of the garbage because I have the Virtual Memory and the Cache for FireFox and various other temp files on a RAM-Disk.

I still count that as 24/7.

I'm running Win7 Ultimate and I have to say that I will not be moving to that piece of garbage Win8.

 

[Ketchup]

.... that piece of garbage Win8.

Shhh, you can't type stuff like that too loud in the forums, lol. At least we're in Mem and Stor, so you may got off with a warning.

BTW, I have used 8 and will be skipping it as well.

When I am at work and have to check on one of our client's Linux boxes, I am always amazed by the uptime. A year is not uncommon.

 

[Deders]

Note the higher latency actually yielded a higher max fps, while the minimums are lower. This is basically saying its all within the margin of error, ie insignificant.

On first glances that makes me think that the ram they tested at 1600 with can't actually handle Cas9 properly and reports more errors that is has to recalculate than it does at cas 10. but then there are so many latency values these days that do make an impact on different tasks It's hard to evaluate properly.

Could just be an anomoly like the one that makes LGA1156 chips overclock better when the fsb multiplier is at odd values, instead of even.

 

[Nec_V20]

Shhh, you can't type stuff like that too loud in the forums, lol. At least we're in Mem and Stor, so you may got off with a warning.

BTW, I have used 8 and will be skipping it as well.

When I am at work and have to check on one of our client's Linux boxes, I am always amazed by the uptime. A year is not uncommon.

When I was Senior German Engineer for Enterprise Disaster-Recovery Tech-Support I talked to one German customer who was worried about his Netware 3 Server and Y2K.

It had been running 24/7 for about 12 years.

To say I was gobsmacked is somewhat of an understatement.

The thing is however with a Win7 PC I do a lot of things with various software and they are not all very tidy. So garbage collects and I have just gotten into the routine of rebooting once a week.

FireFox has a tendency to misbehave after about six days or so. There again I have about 40 tabs open.

I've found that if I treat my Windows installation right and am aware of the limitations then it behaves quite nicely.

I will give Microsoft a pass for Vista because that disaster was more Intel's fault than Microsofts. However Win8 is all home grown. I can't really see Microsoft getting any better until Ballmer is gone. He's never been a techie, just a pen pusher.

 

[Ketchup]

I will give Microsoft a pass for Vista because that disaster was more Intel's fault than Microsofts.

Interesting that you mentioned Vista. I just switched from an XP VM to Vista (for personal stuff on my work laptop), and it is running OK on 2 cores, and 2 GB of RAM. At this point, it's a very 7-like experience. Back when it was released, I think Microsoft really messed up here when they said what it required. Specs were way too low, and combined with the new GIU, people hated it for that. Loaded now with SP2, I can't say a have many qualms with it. Had to disable quite a few things to get it to run well, but it is a VM after all.

 

[Nec_V20]

Interesting that you mentioned Vista. I just switched from an XP VM to Vista (for personal stuff on my work laptop), and it is running OK on 2 cores, and 2 GB of RAM. At this point, it's a very 7-like experience. Back when it was released, I think Microsoft really messed up here when they said what it required. Specs were way too low, and combined with the new GIU, people hated it for that. Loaded now with SP2, I can't say a have many qualms with it. Had to disable quite a few things to get it to run well, but it is a VM after all.

Actually the whole development of Longhorn (codename for what was to become Vista) was premised on Intel going 64-bit with their processors. Intel decided that they would not and the 32-bit compatibility box became the OS.

 

[Nec_V20]

Interesting that you mentioned Vista. I just switched from an XP VM to Vista (for personal stuff on my work laptop), and it is running OK on 2 cores, and 2 GB of RAM. At this point, it's a very 7-like experience. Back when it was released, I think Microsoft really messed up here when they said what it required. Specs were way too low, and combined with the new GIU, people hated it for that. Loaded now with SP2, I can't say a have many qualms with it. Had to disable quite a few things to get it to run well, but it is a VM after all.

Actually the whole development of Longhorn (codename for what was to become Vista) was premised on Intel going 64-bit with their processors. Intel decided that they would not and the 32-bit compatibility box became the OS.

OOPS sorry about the double post

 

[lamedude]

What RAM speed does a NEC V20 need⸮ Being 30% faster than a 8088 I would assume it would need 30% faster RAM.

 

[Nec_V20]

That was my first baby. I upgraded from a 4.77 MHz 8088 to an 8 MHz NEC V20. I had to also solder a new quartz oscillator onto the board.

It still remains a fond and endearing memory as you so astutely noticed.

It was a 67% increase not 30%.

 

[Concillian]

Latency is a one-off penalty. It is a unit of delay that is applied once to every operation. If you are only reading one byte at a time from RAM then yes latency matters. But when you're reading bursts of 1KB or 128KB or more, that tiny bit of latency will not matter. Since every memory read is usually done in large blocks, latency is pretty much meaningless. (Latency is huge for L1 caches, its why intel is so much faster) You can search the web and find several benchamrks where the only thing they change is the CL and it hardly affects anything.

Note the bottom two bars on this chart:

Note the higher latency actually yielded a higher max fps, while the minimums are lower. This is basically saying its all within the margin of error, ie insignificant.

Most important thing to learn about RAM... latency in cycles doesn't matter. Latency in time (or proportional to time) matters.

1066 CAS7 is 7 cycles every 1/1066 seconds (actually double that since it's DDR) = 14/1066 = .0131 seconds

1600 CAS9 is 9 cycles every 1/1600 seconds (times 2 for DDR) = 18/1600 = 0.0113

So while it seems like 1066 CAS7 is lower latency than 1600 CAS9, it's not. The CAS 9 RAM is lower latency AND higher bandwidth. With the DDR3 1600 has a higher CAS number, the actual latency is not higher, but actually lower than CAS7 1066 RAM.

You can't compare CAS latency numbers directly, it tells you nothing. divide the CAS by the advertised frequency and compare that to the CAS / speed of other RAM you're considering.

As a fun aside, I don't know if you were around, but when DDR1-400 RAM was the norm, CAS 2 was considered lightning fast. How much latency is that? 4/200 = 0.020... so really low latency RAM is about half the latency of 10 year old really low latency RAM, but speed / bandwidth has sextupled (2400 vs. 400).

These days however, the impact of both latency and bandwidth is quite small. You can see in that Crysis benchmark that other than running single channel, FPS differences were on the order of 2-5%. Certainly not something to be dropping an extra $100 on in most cases. Blain is pretty much right... get the quality you want and something that will WORK (JEDEC standard voltages). Going beyond that is only something to do if you're squeezing out every last drop for whatever reason.

 

[Z15CAM]

4 X's 4 sticks of Samsung MV-3V4G3D-US_DDR3 in Dual Mode running at 1867 Mhz 9-9-9-24 1T @1.35 volts. It has been very stable and don't have to worry about damaging the CPU's Memory Controller.

I'm running a 4 GB's RAMDisk of it to sluff various Win Read/Writes and App Caching for a Samsung 840 Pro 256 MB SSD.

AIDA64 Extreme Edition Bench

Memory Read: 22127MB/s

Memory Write: 23152 MB/s

Memory Copy: 24206 MB/s

Memory Latency: 39.5 ns

The Samsung 4 GB Green Sticks Scale approx as follows:

- 1600MHz ( 7- 7- 7-20 1T) @ 1.50v WOW! 1600Mhz CL7
- 1600MHz ( 7- 8- 8-24 1T) @ 1.40v
- 1600MHz ( 6- 7- 7-18 1T) @ 1.55v Absolutely Incredable
- 1866MHz ( 9- 9- 9-27 1T) @ 1.40v
- 1866MHz ( 9- 9- 9-24 1T) @ 1.35/1.45v ( I'll take this Setting ;o)
- 1866MHz ( 8- 9- 9-24 1T) @ 1.50v LIKE this CL8 Timming ;o)
- 1866MHz ( 8- 8- 8-24 1T) @ 1.55v Theroetical
- 2000MHz ( 9-10-10-28 1T) @ 1.45v
- 2133MHz (10-10-10-28 1T) @ 1.45v
- 2133MHz ( 9-10-10-24 1T) @ 1.475v
- 2133mhZ ( 9- 9- 9-24 1T) @ 1.50v << Incredable >>
- 2133MHZ ( 9-10-10-28 1T) @ 1.55v-1.6v
- 2133MHz ( 9-11-11-21-1T @ 1.55v
- 2400MHz (11-11-11-28 2T) @ 1.50v

- 2760MHz (11-13-13-35-2T) @ 1.65V

I picked up 16GB's (4 x's 4 GB sticks) of this Samsung Low Voltage, Low Profile, Green Ram around March for $96 delivered from Newegg.ca. They only came in single 4 GB or 2 x's 4 GB stick pkgs. It wasn't on the Market for long - I honestly believe the Gov't and Military bought it all up - I'm Happy.

Platform is an ASUS P8Z68V-Pro Gen3 with an 17 2700K OC'd 1600/4600 Mhz and .996/1.35 Volts 24/7. Temperatures are approx 36/67C with a Corsair H100. C6 is on Auto but I run all 8 Cores with Parking Disabled using a Balanced Win7 [64] Power Profile. Samsung 840 Pro 256 MB, 2 WD5001AALS_HDD's in RAID 0 and 1 WD1002FAEX-0 for Hard Copies and BackUp. Hauppage WinTV 1250 PCIe, Intel Spare NIC PCI Card and a EVGA e-GeForce GTX 280 which I'm about to UpGrade pending what AMD offers (Thinking a GTX 780) to go with a QX 2710 which was damage in shipping waiting replacement. XFX 850W Black Edition Modular PSU. Case is a Fractal Midi ARC 2.

DDR1-400 RAM

I run 3 GB's of OCZ Platinum at 2-3-2-5 IT 408 MHz's in my GA-7N400 Pro 2 (Rev 2) running a Barton 1.833 at 2300+Mhz's, HD 2400 XT with HDMI - been doing it for 10 years now.

How about running a Slot 1 with SD v150 at 138 Mhz in an ASUS P3V4X, VIA Chip Set and a 200Mhz Clock Gen, with a vidded Flip Chip Taulatin 1.3 Celeron at 1800 Mhz's on Air - Still got that HotRod running WinSE with a 9800XT. One of the Fastest Slot 1's in the World I'm told - That one blew the ABIT BH6 outta the water ;o)

 

[aigomorla]

i put a fan over my ram with heat spreaders....
And i think most of the people complaining about not getting enough air to ram, haven't planned out a well designed thermal map.
Also i think they still haven't learned what vendors consider passive means no active fan on the unit.. however a fan must be pointed 5 inches away or ducted like a server.
Vendors definition of PASSIVE!


i dont understand the hate with heat spreaders.
They work if u use them as designed.
They dont work, if u think they can be left passive because as u said, you will just pile up heat in between.
However again... i think you got the definition of Passive wrong... as vendors are using server passive which is ducted.