Tested vs SPD?

[Iamazn]

I am currently looking to purchase this memory kit. There are two sets of specs, the tested and the SPD. What's the difference? If I'm looking for a 2133mhz CL9 memory kit, is this the right kit for me? Are there any better alternatives?
Also, is there any real benefit from 32GB over 16GB of RAM?

 

[larrymoencurly]

Tested = much more lax standards than factory.

You may want to get a 2133MHz kit made with real 2133MHz chips, rather than with 1333MHz chips like these, which were found on some other 2133MHz Ripjaws:

 

[Iamazn]

Tested = much more lax standards than factory.

You may want to get a 2133MHz kit made with real 2133MHz chips, rather than with 1333MHz chips like these, which were found on some other 2133MHz Ripjaws:

Can you recommend a kit that uses 2133MHz chips?

 

[trents]

Why do you want such high speed memory? Have you done any research on the performance gain of high speed memory over 1333 mhz memory. There is little to none. http://www.anandtech.com/show/4503/sandy-bridge-memory-scaling-choosing-the-best-ddr3/6

And concerning the amount, 32 gb vs. 16 gb? Unless you do a lot of Photoshop, CAD, or AV editing/rendering there is no gain with anything over 8 gb, maybe even 4 gb.

 

[Iamazn]

larrymoencurly:
How do I tell if the kit I posted uses 2133MHz chips or 1333MHz chips?

Why do you want such high speed memory? Have you done any research on the performance gain of high speed memory over 1333 mhz memory. There is little to none. http://www.anandtech.com/show/4503/sandy-bridge-memory-scaling-choosing-the-best-ddr3/6

And concerning the amount, 32 gb vs. 16 gb? Unless you do a lot of Photoshop, CAD, or AV editing/rendering there is no gain with anything over 8 gb, maybe even 4 gb.

To be honest, I just want to hit 7.9 on the WEI... And even if the performance gain is little to nothing, it wouldn't hurt going for the little, right? As for the 32GB vs 16GB, I want AT LEAST 8GB RAM in my system. I see using 2x8GB kits as a potential to expand in the future, just in case.

 

[Woomack]

Tested = much more lax standards than factory.

You may want to get a 2133MHz kit made with real 2133MHz chips, rather than with 1333MHz chips like these, which were found on some other 2133MHz Ripjaws:

There is no such thing as 1333 or 2133 chips. All are made whithin some specification and selected by manufacturers for slower or faster kits ( binning ).
Hynix on your photo can be found in every kits from 1333 to 2400 and from batch depends if most of them will make higher or lower clocks. If these chips were in 2133 kit then it means that G.Skill tested it for compatibility and stability on at least 2133 clock.
You have to realise that now on market is almost only Hynix and Samsung and most these chips base on the same architecture.

All memory has SPD settings that are within Jedec specification so in most cases it's 1066 or 1333 9-9-9-24 1.5V +/- 0.05V ( sometimes but really rare can be 1600 9-9-9-24 that can be found in some 2133+ kits ).
Tested settings = what memory manufacturer ( for example G.Skill, Corsair, Kingston, ... ) tested in their labs and most of actually ordered chips reached speed on the label and some more ( usually it's +some % so even worst chips will overclock to some point ).

Windows WEI score is useless number that says you almost nothing about your pc performance. It's something that supposed to give you info if software tested by Microsoft will run without problems on your pc. In real even much lower scores let you use everything.
Memory above 8GB is needed only in some situations , mainly in servers, databases etc. and for home usage you can stop on something like photo/video editing.

I have 7.9 WEI score for my memory in daily pc which is now 6GB DDR3-1680 8-8-8-24 so as you see you don't need anything special to reach max.

 

[Knufire]

. WEI isn't even consistant. I moved the same set of RAM from one system to another, and it dropped from 7.5 to 5.9. Same ram, same timings, same clocks. I moved it back to the original system and it scored 6.someth

You're pretty much throwing your money away. No matter what your use, DDR3-1600 CL9 1.5V is plenty. 8GB if your gaming, 16GB if your doing something that's RAM intensive. If you're buying faster than that and not doing anything that has a tangible benefit from high speed RAM, you're buying it for the wrong reasons. Save the high speed sticks for the people who need it.

 

[Iamazn]

There is no such thing as 1333 or 2133 chips. All are made whithin some specification and selected by manufacturers for slower or faster kits ( binning ).
Hynix on your photo can be found in every kits from 1333 to 2400 and from batch depends if most of them will make higher or lower clocks. If these chips were in 2133 kit then it means that G.Skill tested it for compatibility and stability on at least 2133 clock.
You have to realise that now on market is almost only Hynix and Samsung and most these chips base on the same architecture.

All memory has SPD settings that are within Jedec specification so in most cases it's 1066 or 1333 9-9-9-24 1.5V +/- 0.05V ( sometimes but really rare can be 1600 9-9-9-24 that can be found in some 2133+ kits ).
Tested settings = what memory manufacturer ( for example G.Skill, Corsair, Kingston, ... ) tested in their labs and most of actually ordered chips reached speed on the label and some more ( usually it's +some % so even worst chips will overclock to some point ).

Windows WEI score is useless number that says you almost nothing about your pc performance. It's something that supposed to give you info if software tested by Microsoft will run without problems on your pc. In real even much lower scores let you use everything.
Memory above 8GB is needed only in some situations , mainly in servers, databases etc. and for home usage you can stop on something like photo/video editing.

I have 7.9 WEI score for my memory in daily pc which is now 6GB DDR3-1680 8-8-8-24 so as you see you don't need anything special to reach max.

-Will the stock speed on the kit I posted be the tested specs or SPD specs?
-I know WEI is useless; I just want it maxed for some reason... :/

. WEI isn't even consistant. I moved the same set of RAM from one system to another, and it dropped from 7.5 to 5.9. Same ram, same timings, same clocks. I moved it back to the original system and it scored 6.someth

You're pretty much throwing your money away. No matter what your use, DDR3-1600 CL9 1.5V is plenty. 8GB if your gaming, 16GB if your doing something that's RAM intensive. If you're buying faster than that and not doing anything that has a tangible benefit from high speed RAM, you're buying it for the wrong reasons. Save the high speed sticks for the people who need it.

My main purpose will be gaming, but I will also be hosting a server for a few friends to play on.

 

[Woomack]

-Will the stock speed on the kit I posted be the tested specs or SPD specs?
-I know WEI is useless; I just want it maxed for some reason... :/

All memory after 1st start will use SPD settings ( probably 1333 9-9-9-24 1.5V ). After that you have switch to XMP profile in BIOS that will include 2133 settings.
It's because not all boards can run at so high settings and Jedec specs for SPD should be available for all.

My main purpose will be gaming, but I will also be hosting a server for a few friends to play on.

Ram in link should be fine for everything mentioned.

 

[larrymoencurly]

G.Skill RipJawsX 2133 MHz made with 1333 MHz chips:

There is no such thing as 1333 or 2133 chips. All are made whithin some specification and selected by manufacturers for slower or faster kits ( binning ).
Hynix on your photo can be found in every kits from 1333 to 2400 and from batch depends if most of them will make higher or lower clocks.

Hynix DDR3 H5TQ2G83BFR data sheet

From the Hynix part number decoder:

SPEED(tCL-tRCD-tRP)

TE: DDR3-2133 14-14-14
RD: DDR3-1866 13-13-13
PB: DDR3-1600 11-11-11
H9: DDR3-1333 9-9-9
G7: DDR3-1066 7-7-7
S6: DDR3-800 6-6-6

Those chips are binned at CLK9 1333 MHz. So how can you say there's no such things as 1333 MHz or 2133 MHz chips? And because Hynix rated those chips at 1333 MHz, I don't see how they can be legitimately used in 2400 MHz modules, except in the land of "everybody does it" and "good enough".

If these chips were in 2133 kit then it means that G.Skill tested it for compatibility and stability on at least 2133 clock.

I mean legitimate binning, not something done with just MemTest86+ in PCs running the memory at way below the chips' maximum rating of 85C, like 25C - 58C.

 

[larrymoencurly]

larrymoencurly:
How do I tell if the kit I posted uses 2133MHz chips or 1333MHz chips?

Read the writing on the chips, without even opening the retail packaging. This can almost always be done with first-rate modules, as opposed to those modules decorated with pec implants and cod pieces heat spreaders.

From a practical standpoint, buy Samsung or no-heatsink Crucial modules. I don't see any Samsung made with chips rated faster than 1600 MHz (PC12800), but it's known to be very overclockable and usually able to run faster than modules rated for 2133 MHz, probably it's made with chips that aren't already overclocked:

http://www.newegg.com/Product/Product.aspx?Item=N82E16820147096

Don't worry about the 1.35V rating because the chips are rated to be comptible with 1.5V operation and have the same absolute maximum rating of 1.975V.

 

[Woomack]

... because there isn't such thing as 1333 or 2133 chips , as you already said all are binned and then marked by IC manufacturers. H9 is most popular series in highest memory kits so 2400-3000 that is working on medium timings with high clock. Other way you can read it as 2133 CL14 won't make 2600 CL12 and 1333 CL9 can make it.
If these 2133 CL14 were much better then memory manufacturers would make modules on these chips.
Noone is testing it with memtest. They have hardware testers.

 

[Iamazn]

Ram in link should be fine for everything mentioned.

Would it be possible to overclock this kit even further?

 

[Woomack]

Would it be possible to overclock this kit even further?

Probably to something like 2400 10-12-11 but you won't see any difference in performance so there is no point.

 

[larrymoencurly]

... because there isn't such thing as 1333 or 2133 chips , as you already said all are binned and then marked by IC manufacturers. H9 is most popular series in highest memory kits so 2400-3000 that is working on medium timings with high clock. Other way you can read it as 2133 CL14 won't make 2600 CL12 and 1333 CL9 can make it.
If these 2133 CL14 were much better then memory manufacturers would make modules on these chips.

I said there are 1333 MHz and 2133 MHz DDR3 chips because Hynix said they made them. Did you even look at the data sheet I linked? But by your line of reasoning, there are also no such things as 5% tolerance resistors or 10% tolerance resistors because they're merely binned to those rating, but it's binning that gives parts their ratings, so you're being pedantic.

Noone is testing it with memtest. They have hardware testers.

I never said they used memtest. I did say they used MemTest86/86+, but what's the difference to anybody who thinks 1333 MHz and 2133 MHz chips don't exist.

A factory tour of Kingmax showed that they had an Advantest machine. A YouTube video of a Kingston production line also revealed one, and several years ago Kingston said it had several but didn't use them for all models or for all OEM or industrial customers. Geil and G.Skill showed only PC-based testing and told me by e-mail that they used MemTest86+. Factory tours of Patriot and Corsair also showed only PC-based testing, and Micron/Spectek's own video revealed no expensive hardware testers for Spectek chips. I couldn't find information for OCZ or Mushkin, but I seriously doubt all their DRAM is tested with expensive machines. Many companies have automatic loaders and ovens, but they're still PC-based testers, without any special high-speed, random logic-based testing hardware.

 

[Iamazn]

I said there are 1333 MHz and 2133 MHz DDR3 chips because Hynix said they made them. Did you even look at the data sheet I linked? But by your line of reasoning, there are also no such things as 5% tolerance resistors or 10% tolerance resistors because they're merely binned to those rating, but it's binning that gives parts their ratings, so you're being pedantic.


I never said they used memtest. I did say they used MemTest86/86+, but what's the difference to anybody who thinks 1333 MHz and 2133 MHz chips don't exist.

A factory tour of Kingmax showed that they had an Advantest machine. A YouTube video of a Kingston production line also revealed one, and several years ago Kingston said it had several but didn't use them for all models or for all OEM or industrial customers. Geil and G.Skill showed only PC-based testing and told me by e-mail that they used MemTest86+. Factory tours of Patriot and Corsair also showed only PC-based testing, and Micron/Spectek's own video revealed no expensive hardware testers for Spectek chips. I couldn't find information for OCZ or Mushkin, but I seriously doubt all their DRAM is tested with expensive machines. Many companies have automatic loaders and ovens, but they're still PC-based testers, without any special high-speed, random logic-based testing hardware.

Then what kit should I buy if I'm looking for:
-2133mhz CL9 2x8GB?
-2133mhz CL9 2x4GB?
Also, how would this kit:
http://www.amazon.com/gp/product/B0...&pf_rd_t=101&pf_rd_p=470938631&pf_rd_i=507846
compare with the kit I originally posted?

 

[larrymoencurly]

Then what kit should I buy if I'm looking for:
-2133mhz CL9 2x8GB?
-2133mhz CL9 2x4GB?
Also, how would this kit:
http://www.amazon.com/gp/product/B0...&pf_rd_t=101&pf_rd_p=470938631&pf_rd_i=507846
compare with the kit I originally posted?

Something rated for no more than 1.5V at its fastest speed, with no heatsinks and with the fastest rated name brand chips, even if the ratings for those name brand chips are slower than your desired 2133 MHz.

I'd really favor those 1.35V Samsungs because of their reputation for being so overclockable

Check the clearance with your CPU heatsink if you get any DIMMs with heatsinks that stick up.

 

[Iamazn]

Something rated for no more than 1.5V at its fastest speed, with no heatsinks and with the fastest rated name brand chips, even if the ratings for those name brand chips are slower than your desired 2133 MHz.

I'd really favor those 1.35V Samsungs because of their reputation for being so overclockable

Check the clearance with your CPU heatsink if you get any DIMMs with heatsinks that stick up.

I sorta want the speed to be 2133mhz at stock, so I don't have to do any manual overclocking...

 

[larrymoencurly]

I sorta want the speed to be 2133mhz at stock, so I don't have to do any manual overclocking...

In practical terms, considering what kind of memory available on the maket, your only choices are:

A. Good manual overclocking (run Samsung 1600 MHz @ 2133 MHz)

B. Bad automatic overclocking (G. Skill 1333 MHz @ 2133 MHz)

C. Manually under-overclocking the automatically overclocked memory (G. Skill 1333 MHz @ 1333 MHz)

If I had to overclock, I'd choose A first, no questions, especially because it has such a good track record, and 1600 MHz chips can probably run faster than 1333 MHz chips. And with those Samsung modules you'll get something that's rock-solid at 1600 MHz, no matter which sample you receive.

 

[hokiealumnus]

What on earth are you talking about larrymoencurly? G.Skill rates their sticks for what they will run. I have never had a set of G.Skill RAM fail to run its rated speed unless it was an IMC problem, from DDR2-1066 to DDR3-2666. However there is a strong potential for a Samsung 1600 kit to not run at 2133. Plus, because it's not rated for where he wants to run, he's out of luck on an RMA.