Best Brand

xCWolf

[H]ard|Gawd
Joined
Aug 12, 2011
Messages
1,521
What are your opinions on best brand for DDR3. I've always been a corsair fan myself but from the sigs i've read around here it seems like there are alot of OCZ fans.

so whats your favorite RAM and why?
 
Corsair(Non vengance) is my number 1. My alternative is Mushkin. I do not go outside of those 2.
 
I wouldn't say that there's a lot of OCZ fans, they just had cheap alternatives at the time. :p

That being said, I've been using G.Skill for quite awhile now without any ill things to report. I've had to RMA a kit to them which was absolutely painless. Their forums are always crawling with people and they respond to emails extremely quickly. Not to mention, I keep seeing their kits consistently cheaper than equivalent kits from others.

Mushkin is also another go-to for me as I have yet to have any issues with RAM I've bought from them.

Crucial is another story (damn you DDR2 BallistiX).
 
I LOVE patriot. Rock solid, can typically go beyond their rated specs, good warranty, and they look good in a windowed case. All around the best ram I have ever used. Benches better than most ram as well ;)

I however, am currently using Corsair dominator GT's because I have airflow problems, and wanted to have a ram fan and i wasnt sure if it would fit over my patriots, so i sold the patriots and got the corsair GT's. They're pretty, and they do their job, but XMP did not work out of the box, and they were finicky to stabilize - i have higher voltages across the board in comparison to what i had with patriots.

I would consider mushkin and gskill because of their low published failure rate, but that wont happen until i run into some bad patriot vipers, or bad corsair dominators.

I don't like OCZ ram because they failed hard in my ex's computer, and she wasnt even overclocked - i dont mind at all that they switched over to SSDs hahaa.
 
What are your opinions on best brand for DDR3. I've always been a corsair fan myself but from the sigs i've read around here it seems like there are alot of OCZ fans.

so whats your favorite RAM and why?

Best performing or best value? I had G.Skill and running Corsair now. It's really comes down for me - what's the best value from known brand which delivers specs x and y at best price. I never look for motherboard suggested kits
 
Where I work, we go off reliability as the #1 factor when selecting RAM. Based on the three main brands we've tried, for long periods, over the years I have to say Kingston is the best. Half the failure rate of Corsair, and about 1/10th the problem rate of OCZ.

Now that is their ValueRAM line, which I think is ideal if you aren't overclocking. We have had issues with HyperX sometimes, and are now carrying Patriot on the higher end. Haven't been carrying it long enough to have statistically worthwhile data, though.
 
i've read around here it seems like there are alot of OCZ fans.

Maybe 3 years ago I would have included them in my list of memory manufacturers to buy from but now I would not buy anything at all from them (SSDs, memory or power supplies).

My current best picks (at home and work) are Corsair and GSkill. I am still staying away from Patriot since these caused a lot of problems in the past with AMD boards. Some times I will pick up a crucial or kingston however I prefer GSkill and Corsair.
 
When I need memory, I'll shop G.Skill, Mushkin, Corsair, and Crucial. I buy what's cheapest for the specs / capacity I'm looking for at the time. Sometimes you can find stuff on sale at really good prices. If the cost were the same for them all, I'd probably buy Mushkin.

I've got Mushkin in my #1 rig at home and my workstation at work (yes, I got to build my own workstation) and G.Skill in the wife's machine at home. Never had a problem with any of them.
 
G.Skill, Mushkin, and Crucial for me.

My favorite is G.Skill as I had a pair freeze on me and the RMA process was a piece of cake. Made me like them even more.
 
G Skill is great. I've had several of their kits, never any problems.

Great warranty too. From why I've heard, they are very good at RMAing if ever a problem occurs.
 
I thought OCZ didn't make ram aymore.
Anyways I have OCZ 1600 Platinum and I don't like it :)
It was troublesome figuring out how to run it at it's rated speed.

Next time around I'd go corsair with XMP on the chip.
I feel like buying their 12gb triple channel kit...but maybe it will go on sale :)
 
So it looks like the Corsair G.Skill or Mushkin are the brands to go with!

Thanks Guys
 
Kingston are good bottom line any memory purchases just make sure you get the "lifetime warranty" in case of any issues.

It's not the brand that matters there are tons of so called memory brands out there sprung up from nowhere it seems. It's the chips (and quality) they use that does matter.

Crucial well I've had a few too many returns last few years and the last RMA they took weeks to get it back to me so I'm not using them again. Awful customer service. (in the UK anyway)

And they thought it was OK to replace expensive ballistix with normal ram I lost quite a bit of cash on that I can tell you.
 
It's not the brand that matters there are tons of so called memory brands out there sprung up from nowhere it seems. It's the chips (and quality) they use that does matter.
...yes and no.

Yes, the memory chips and the quality of them makes a huge difference. But, each company designs their own module (circuit board) and let me assure you, high speed signal integrity on that module can make or break the memory performance.

Things like layer count on the PC board, PC board materials, signal routing, controlled impedance, etc. are all huge factors in how well the memory performs. Add to that how well the module is built (soldering, etc.) and there can be dramatic differences in how identical chips perform on different modules.

So, at the end of the day, the company you buy from can make a huge difference.
 
...yes and no.


So, at the end of the day, the company you buy from can make a huge difference.

Not really because memory performance is one of the least useful areas to tinker with. Bang per buck it's not even on the chart and this has been shown through the years. Even moving from DDR2 up to DDR3 didn't do a lot. Faster modules don't really show any significant peformance increase it's pretty minor at best.

If you want more fps you get a better graphics card if you want more CPU performance overclock it or get a faster one both are far more useful than memory. It's one of the biggest myths with pc's or course makers want you to believe their nice premium ram is so much faster..it's not..it's a small improvement at best.
 
Not really because memory performance is one of the least useful areas to tinker with. Bang per buck it's not even on the chart and this has been shown through the years. Even moving from DDR2 up to DDR3 didn't do a lot. Faster modules don't really show any significant peformance increase it's pretty minor at best.

If you want more fps you get a better graphics card if you want more CPU performance overclock it or get a faster one both are far more useful than memory. It's one of the biggest myths with pc's or course makers want you to believe their nice premium ram is so much faster..it's not..it's a small improvement at best.
I'm not referring to speed, I'm talking about stability and reliability.

You are correct; there is of course RAM with slightly faster timings, etc. and I'll concur that the performance gains aren't what a new video card brings to the table. Just so happens that lots of that RAM also has a better engineered module, with better PCB layouts and better PCB materials that improve signal integrity of memory bus.

I've had systems on my bench that failed MemTest86+ badly.....and moving the modules around from slot to slot fixed the problem. That's an example of memory that is on the edge, timing and signal integrity wise. I don't seem to see those problems with performance memory nearly as much.

With any new system or memory upgrade, testing really needs to be done to verify stability. At a minimum, systems that come to me run MemTest86+ for 24 hours. At the end of the test, "No Errors" is the only acceptable answer. You would be surprised how many systems, even some purchased as pre-built systems from big names, won't pass that test.
 
I think I'm just going to buy some mushkin =P looks cool and sounds like it works!
 
Ignore all the wrong answers. :)

What really matters is that the chips on the module are clearly identified as being from Micron, Samsung (SEC), Hynix, Winbond, ProMOS, PowerChip, Nanya/Inotera, Elpida, etc., rather than having a private label or even no markings at all. "Clearly identified" means you don't have to remove a heatsink to check the brand.
 
Ignore all the wrong answers. :)

What really matters is that the chips on the module are clearly identified as being from Micron, Samsung (SEC), Hynix, Winbond, ProMOS, PowerChip, Nanya/Inotera, Elpida, etc., rather than having a private label or even no markings at all. "Clearly identified" means you don't have to remove a heatsink to check the brand.
....I'll bite - you don't think the module PCB, it's construction, layer count, signal integrety, etc. make any difference at all...??
 
Samsung started producing its own low profile DIMM @ 1.5v. I would have bought that if not already owning my current set.
 
Ignore all the wrong answers. :)

What really matters is that the chips on the module are clearly identified as being from Micron, Samsung (SEC), Hynix, Winbond, ProMOS, PowerChip, Nanya/Inotera, Elpida, etc., rather than having a private label or even no markings at all. "Clearly identified" means you don't have to remove a heatsink to check the brand.

I just pulled a heatspreader off a corsair stick and there is no brand name on the chip. Going by what you just said this memory isnt good quality?
 
larrymoencurly said:
What really matters is that the chips on the module are clearly identified as being from Micron, Samsung (SEC), Hynix, Winbond, ProMOS, PowerChip, Nanya/Inotera, Elpida, etc., rather than having a private label or even no markings at all. "Clearly identified" means you don't have to remove a heatsink to check the brand.
....I'll bite - you don't think the module PCB, it's construction, layer count, signal integrety, etc. make any difference at all...??
From what you've seen of signal intergrity tests, how much did PCBs differ, net of the DRAM chips, both for modules made with no-name chips and with major brand chips?

I thought everybody's PCBs were either JEDEC reference design, BrainPower, or in the case of chip makers and maybe Kingston, their own designs. I couldn't tell who made the PCBs for Crucial modules containing Micron chips, but I have some Crucials with Samsung chips, and they were made with Samsung PCBs. The Corsair 1GB PC3200 ValueRAMs I bought this year all have BrainPower PCBs (no-name chips, 1 module failed MemTest86), as does a 2GB PC5300 Patriot from Dec. (Hynix chips).
 
Last edited:
I just pulled a heatspreader off a corsair stick and there is no brand name on the chip. Going by what you just said this memory isnt good quality?
Let's say is that it wasn't intentionally made to be of the best quality, and those chips were bought either as factory rejects (euphemistically called "UTT" -- UnTesTed) or as whole, uncut wafers that weren't tested completely at the factory. And if you have Corsair XMS memory, Corsair admits that it's overclocked and not tested very much (they virtually admit so in their description of its "thorough" testing). You should test with at least MemTest86/MemTest86+ and Gold Memory, running each diagnostic for at least 12 hours, preferrably on more than one motherboard and with each module in each DIMM socket. But some mobos, like BioStars, may not allow good testing because they not only default to higher than normal voltage but may not even let you manually set the voltage to normal.

I think my overall failure rate for memory modules is around 10%, but only one of those failures involved name brand chips, and that was with a module with solder slopped over the gold fingers, and I didn't even bother testing it. However I never tried Crucial Ballistix memory.

If the heatspreader was glued on rather than clamped on, I hope you pulled it off by prying between the chip and heatspreader, rather than between the circuit board and heat spreader, which can break solder connections.
 
G. Skill has served me well. I've got 3 systems with ripjaws, and they are all overclocked and stable for 1+ years without issue. Used kingston in my family's boxes, and there hasn't been a failure.

The price for g skill has also been a big seller for me.
 
Let's say is that it wasn't intentionally made to be of the best quality, and those chips were bought either as factory rejects (euphemistically called "UTT" -- UnTesTed) or as whole, uncut wafers that weren't tested completely at the factory. And if you have Corsair XMS memory, Corsair admits that it's overclocked and not tested very much (they virtually admit so in their description of its "thorough" testing). You should test with at least MemTest86/MemTest86+ and Gold Memory, running each diagnostic for at least 12 hours, preferrably on more than one motherboard and with each module in each DIMM socket. But some mobos, like BioStars, may not allow good testing because they not only default to higher than normal voltage but may not even let you manually set the voltage to normal.

I think my overall failure rate for memory modules is around 10%, but only one of those failures involved name brand chips, and that was with a module with solder slopped over the gold fingers, and I didn't even bother testing it. However I never tried Crucial Ballistix memory.

If the heatspreader was glued on rather than clamped on, I hope you pulled it off by prying between the chip and heatspreader, rather than between the circuit board and heat spreader, which can break solder connections.

LOL
 
For what it is worth, here are the statistics for RAM module failures from the company I work with over the last 10 years. Please note that this includes modules that were DOA or found to be defective in our testing here, as well as those which failed later on for customers in the field (if they reported them to us, or were found as part of troubleshooting we were involved with). I am only listing brands for which we have at multiple thousands of modules purchased, to avoid statistically insignificant results, and this is only by brand of module not by chip or PCB maker:

Corsair: 8.53%
Kingston: 2.70%
OCZ: 10.18%

Now obviously we don't have this same volume of statistics for the many other brands out there, so I can't speak to how they may stack up... but from these three brands it has been clear that one was far better for reliability, and that is what we use currently.

If you break it down further, though, this gets interesting:

Kingston ValueRAM: 0.91%
Kingston HyperX: 10.29%

It was the above numbers that caused is to keep using Kingston's standard line of RAM, but switch to Patriot for our performance memory. We haven't used them long enough for me to have any significant report on overall quality and reliability, though... but I have high hopes :)
 
G.Skill has been the least problematic for me. I also like Corsair, Crucial, Kingston, and Mushkin.
The OCZ sticks I purchased are perhaps the most finicky sticks I've ever come across, and regret buying them because of that and the fact that super tall heat sinks are pretty much useless lol.
 
I personally go for Kingston, even if the price is a bit higher as I really value their reliability and compatibility . If their price is a lot higher, I probably wouldn't mind going for AData, Mushkin, G.Skill, or Wintec (haven't had problems with them to date).

I can tell you definitively who I would avoid like the plague though: OCZ. Nothing but problems with them.
 
I've had good and bad Kingston and OCZ, nothing but good from G.Skill. I'll stay with G.Skill until they do something foolish and make me change; I'm not even really enticed by the other stuff (except when they put LEDs on the top... lol)
 
Mushkin assembled and supported right here in the USA. There is not any other brand I will use.

When the first round of Intel I7/I5's came out I went through 4 kits of ram before I tried Mushkin, ranging from G-Skill, Supertalent, Corsair and Kingston. I was shooting for the lower latency higher XMP clocked ram. All those other kits failed, could not run flash with any of them installed. Tried Mushkin on a whim and have been sold ever since. They also have excellent support if I did not mention that already.
 
Crucial is also an American company, most of their actual chips (Micron) are manufactured in Boise.

Not only is Mushkin an American company, they're an employee owned cooperative. Happy workers = good products & good service, and they're price competitive.
 
Ah yeah, I had forgotten about Crucial. I think I've put them in someone's computer... oh, that's right I had to buy a stick of them when I needed a DDR3 stick fast. Yeah I'd definitely buy them again.

Also, I did not know that mushkin was located in the USA. That's pretty cool.
 
Back
Top