QVL - Myth, Legend, Marketing/Advertising, what is your take?

[sneekypeet]

Through a couple of previous threads, a discussion popped up about the value of a RAM QVL from either the RAM makers or the motherboard makers when building a system.

Rather than to clog up other threads, lets have an educated conversation and see of we can come to a consensus of its merrits.



One side of the coin is it is the definite guide to what RAM to pick, whether that specific model number, or one with the same specs from another maker.

Another side is that the QVL is more or less a list of RAM that the big boys in the industry push to the world to ensure they are listed on a guide, more as advertising, less as an important guide.

As always, parts choice does matter. It is not as broad as I would like it to be saying QVL is completely useless. I also added a poll.

 

[freeagent]

I used to hold no value in it, and I had problems occasionally.

The one time I adhered to the QVL (this system), smooth sailing, no problems ever..

 

[Macro Device]

QVL is clearly not any exhausting, it's just a list of RAM sticks the manufacturer cared to test and confirm it works as intended. How much marketing and conspiracy there is, not for me to judge, no idea here. I personally mostly ignore this list because I know I can afford a couple circles around the block to buy a different set of RAM in case things go south. Never happened to be the case though. All RAM I put into my boards, even the most no-name knock off turdstones, worked no less than they should have. That could be my luck though.

So I'd call it... "QVL is a tour guide for those who have absolute void for ideas what they are doing."

 

[sneekypeet]

QVL is clearly not any exhausting, it's just a list of RAM sticks the manufacturer cared to test and confirm it works as intended. How much marketing and conspiracy there is, not for me to judge, no idea here. I personally mostly ignore this list because I know I can afford a couple circles around the block to buy a different set of RAM in case things go south. Never happened to be the case though. All RAM I put into my boards, even the most no-name knock off turdstones, worked no less than they should have. That could be my luck though.

So I'd call it... "QVL is a tour guide for those who have absolute void for ideas what they are doing."

That is a great way of putting it!

 

[Solaris17]

I'm not sure how to vote so I put "it helps". QVL is tested true, on the board. Now does that mean your sticks are going to do X,xxxmhz on the same board? I dont think so. I think you are still fighting with lottery; the dimms themselves, the cpu, the mobo, the driver strength, the memory controller.

Finally, the BIGGEST reason I think ram incompatibility happens. Sub timings. Everyone knows the big 4 maybe 5 and the CR. Some know a small subset of secondaries. However when you start getting super deep into tREFI etc and most people are lost. I think these deeper sub timings are really the devil in the details, but on the flip I think those timings are put as safe as possible by most boards.

I think for the most part anything will work as long as you follow JDEC and I dont mean speed, make sure the primary timings are right. Any failure afterwards from my experience are the sub timings or any XMP/EXPO profile that touches them.

 

[Macro Device]

I think you are still fighting with lottery; the dimms themselves, the cpu, the mobo, the driver strength, the memory controller.

Don't forget the waves. I had my PC performance cut in half by some guy testing his makeshift reactor a couple stories above me. Luckily they arrested him...

 

[ir_cow]

I should open a business where you send me a memory kit not in the QVL list and if it doesn't work I pay you. If it did you pay me. I would be so rich.

 

[freeagent]

I should open a business where you send me a memory kit not in the QVL list and if it doesn't work I pay you. If it did you pay me. I would be so rich.

I have some Adata's here if you want

 

[damric]

Useless

 

[freeagent]

Useless

Right? If you buy sticks using the most popular chips, but try something off the beaten path

The Adatas I mentioned do work at DOCP. The problem with them is when you kill power to the system, it retains nothing.

Also, no overclocking.. the system literally shits its pants and plays with it.

I have a few different sets of B-Die now and they all work just fine..

 

[eidairaman1]

Through a couple of previous threads, a discussion popped up about the value of a RAM QVL from either the RAM makers or the motherboard makers when building a system.

Rather than to clog up other threads, lets have an educated conversation and see of we can come to a consensus of its merrits.



One side of the coin is it is the definite guide to what RAM to pick, whether that specific model number, or one with the same specs from another maker.

Another side is that the QVL is more or less a list of RAM that the big boys in the industry push to the world to ensure they are listed on a guide, more as advertising, less as an important guide.

As always, parts choice does matter. It is not as broad as I would like it to be saying QVL is completely useless. I also added a poll.

It's to be taken as a minimum essential listing guide of what has been tested by the board makers, it doesn't mean other modules outside of that listing won't work though.

My Ripjaws weren't on the list and they are overclocked to tighter than what Tridents were (Sig Rig)

The 5800 rig I built in 21 uses Ballistix Gamer 3600 on a B550 Steel Legend, and I don't recall if they were on the list either.

Now in the day I had trouble with Kingston and Corsair.

I avoided them during the 5800 Build from users coming in with issues that couldn't be resolved by even using JEDEC for DDR4.

I might give Patriot, Geil, Teamgroup a go if they have non RGB modules (Too many lights are a headache)

My good experience was with Crucial and Mushkin long ago, seem to be pretty consistent.

 

[ShrimpBrime]

Old hat. People don't read the motherboard manual.

So you expect them to read a qvl and understand all the acronyms?

1R and 2R confuses people while they mistake this for memory channels.

___

There are more reasons not to utilize QVL than to guide a new user into looking at it.

I follow Woomack at OCF and IR_Cow at TPU for their memory testing adventures. Many of the kits they test are not on qvl, but geared towards performance. And in most TECH forums, people look for performance. In example, how can I lower my latency? Or can I get the memory to go faster? Funny enough, the forum here also has a large benchmark section that has thousands and thousands of posts. Most of these people seem to be performance users.

A basic user, has an HP or DELL and don't game or need performance. That would be my mother in law. She does here taxes and looks at JC Penny website. She wouldn't even care to know qvl lists. The pc just turns on and works. And does not visit PC TECH forums.

The inaccurate poll should be :

Did you use qvl to buy your ram
Or
Did you NOT use qvl to buy your ram.

And would very easily settle an argument. The current poll simply carries on the argument. (No offence intended)

 

[Contra]

I will say this: QVL is very useful if the board manufacturer has tested a sufficient number of different memory chips in different combinations, and your supposed combination (capacity, modules per channel, speed) is among them. All this gives additional hope when overclocking and indirectly speaks about the capabilities of the board.
For example, if the manufacturer has not tested a single module above 7000, you should not count on 8000, or there is not a single 4x kit, then overclocking such a configuration will not work adequately, because even the board manufacturer does not believe in it).
More often there is a situation when the memory is tested only when the platform is launched, for example for z690 boards in qvl you can see only 8-16GB, but not 24/32/48. This indirectly indicates that the BIOS is not optimized for newer modules

 

[ShrimpBrime]

I will say this: QVL is quite useful if the board manufacturer has tested a sufficient number of different memory chips in different combinations, and your intended combination (capacity, modules per channel, speed) is among them. All this gives additional hope when overclocking.

What about the next 5 months of bios revisions. Do ALL MB manufacturers keep up with testing and updating the lists?

I was looking at a qvl today for an older 6th gen Intel board. It looked very small for testing. Maybe 4 different sets of corsair and many boxes where the die maker goes was blank. So you don't even know what memory chips are installed. And ddr4 has at least 3 common PCB layouts they might use the same part number for.

Overclocking comes from experience and deep knowledge of timings and their purposes. Average overclocker sets only primary timings, all else auto. Experienced, like Igor, a name you all know could probably tune any memory kit out there and see fantastic results. Keeping in mind there's a line between overclocking and extremeoverclocking. But this is all a totally different subject.

So I ask you sir user. Did you use the qvl to purchase your memory kit? Yes or no?

Also and in addition to,

I'm directly going to use the word qvl useless.

Saying used less than more. So it's useless literature most people, even average HP desktop users do Not view and understand.

 

[Contra]

What about the next 5 months of bios revisions. Do ALL MB manufacturers keep up with testing and updating the lists?

I've added a bit to the answer and will add here as well.
QVL can be useful for understanding the potential of a motherboard and I usually look at it, not for buying RAM but for buying a motherboard) Regular users won't do this when buying pre-built PCs.

 

[ir_cow]

Finally, the BIGGEST reason I think ram incompatibility happens. Sub timings. Everyone knows the big 4 maybe 5 and the CR. Some know a small subset of secondaries. However when you start getting super deep into tREFI etc and most people are lost. I think these deeper sub timings are really the devil in the details, but on the flip I think those timings are put as safe as possible by most boards.

I partially agree with this. XMP and EXPO only has a number of values stored, they rest comes from the Motherboard and whatever the BIOS team impalements. I noticed that often its just a linear value. so if 6400 to 7200 is 12%, X value for 6400 is now 12% higher when picking 7200. This method works to a point, but with a bunch of different memory manufactures (Hynix, Micron, Samsung, etc), they don't all do well with linear progression. tWRWR_sg is a big one and very different per brand. So if your stuck in a OC, just raise that value, it might magically start working

XMP: CL-RCD-RP-RAW-RC-RFC1-RFC2-RFCSB-WB (9)
EXPO: CL-RCD-RP-RAW-RC-RFC1-RFC2-RFCSB-WB-tCCD-tRRD-tWTR-tRTP-tFAW-tCCDW (15)

 

[sneekypeet]

@Bill_Bright

Sorry I didnt tag you earlier, but I feel your questions from the other thread would be better taken up here. if you want to copy/paste you last questions I will answer, but I didnt feel right dragging it over without your permission.

 

[Ruru]

QVL is simply a list of RAM/M.2 SSD/GPUs the manufacturer has tested the motherboard with and confirmed to work with it, it's not a "only these listed ones will work" list. I've never gone with QVL what it comes to RAM and haven't had a SINGLE issue since the SDR era.

 

[windwhirl]

Through a couple of previous threads, a discussion popped up about the value of a RAM QVL from either the RAM makers or the motherboard makers when building a system.

Rather than to clog up other threads, lets have an educated conversation and see of we can come to a consensus of its merrits.



One side of the coin is it is the definite guide to what RAM to pick, whether that specific model number, or one with the same specs from another maker.

Another side is that the QVL is more or less a list of RAM that the big boys in the industry push to the world to ensure they are listed on a guide, more as advertising, less as an important guide.

As always, parts choice does matter. It is not as broad as I would like it to be saying QVL is completely useless. I also added a poll.

The list helps but is by no means exhaustive. I live in a place where 99% of QVL memory kits don't exist in the market so, my take is, if you can find one, awesome, if not, you'll just have to get something else

 

[Waldorf]

while i check it, especially if the pc isnt for me, or uses parts i havent used (much), but tired of ppl seeing it as a bible, or dismiss certain brands,
just because they tend to swap chips around (e.g. corsair and others), which really isnt an issues with most ppl only interested in AMP/XMP perf.


@Solaris17
not even if you only switch primaries.
for the first x570 i bought, everyone recommended (QVL) Gskill kit 3600C16, never worked past jedec, even if i used manual entries for voltages/main timings only (to get it to work),
but as soon as i rebooted, it didnt post.
the corsair vengeance kit that everyone said was junk 3600C18, work with the same settings after forced shutdown and just swapping the kit.

@Contra
except many boards will have (multiple) updates by the time most compare/buy, and they might not care to update anything,
meaning you might dismiss one model, just because its info wasnt updated at the time you checked.
so far, all ryzen builds i did since they came out, had no trouble running anything with "known" quality dies (say 3600C16/4000C20; or "better").

 

[JustBenching]

Through a couple of previous threads, a discussion popped up about the value of a RAM QVL from either the RAM makers or the motherboard makers when building a system.

Rather than to clog up other threads, lets have an educated conversation and see of we can come to a consensus of its merrits.



One side of the coin is it is the definite guide to what RAM to pick, whether that specific model number, or one with the same specs from another maker.

Another side is that the QVL is more or less a list of RAM that the big boys in the industry push to the world to ensure they are listed on a guide, more as advertising, less as an important guide.

As always, parts choice does matter. It is not as broad as I would like it to be saying QVL is completely useless. I also added a poll.

QVL is a list of the ram tested with that specific motherboard and work at XMP. No more no less. I never check it cause I'm not running xmp anyways but it's good to have for the people that do.

 

[Waldorf]

still had to many times where stuff wouldnt work at all/not with main timings (as sold), and some mistake tested for "guaranteed".

 

[sneekypeet]

work at XMP

This may not mean what you think it does.

What is that metric, Windows loading? memtest? There is no answer, nor are the supported voltages shown, VDIMM is most likely a given. passing qualification for QVL is different to everyone making those lists.

 

[agent_x007]

Only JEDEC must work, "QVL" is irrelevant. DDR5 RAM must work in DDR5 board - that's the whole point of having a standard.
If it doesn't, a simple BIOS/UEFI patch should be the only thing required to make it work at JEDEC speed (assuming, it's not a hardware limitation).

XMP/EXPO/EPP/etc. are NOT and never will be guaranteed to work.
You are buying possibility of running of that speed (guaranteed by memory manufacturer), BUT that does NOT mean it will work at that speed on MB/platform you own.

QVLs are only required to have something for marketing to based it's pretty numbers on.
If memory maker makes a future batch of "QVL" memory with different chips, and you can throw that "QVL approve" mark/entry into bin (or wait before purchase for new QVL list to be made).

 

[sneekypeet]

QVLs are only required to have something for marketing to based it's pretty numbers on.

I also implied ROI on the RAM makers to submit to be seen, and then be sold, in one of the threads that prompted this