Does memory brand matter alot?

Does memory brand matter when all the spec of the memory would be the same like amount of MHz and DDR?
For instance for assembling a Clevo I´ve seen a brand-less 8Gb ddr3 1600 stick and a 8Gb ddr3 1600 Crucial stating "extreme performance first class"! Is this a just marketing trick or does the 2nd one really be faster?
And does it matter in speed much if I put one stick of 8Gb in it compared to 2 sticks of 4GB so in dual?

 

No, not really. It's mostly just marketing buzzwords.

 

same, what matters is the speed and timings

 

Brand usually doesn't matter, the specs (speed, CAS latency, etc) are what makes the difference. That said, some laptops may be a bit picky about RAM, where one brand will work and another will cause some issues--but that's the exception to the rule.

As for 2x4GB or 1x8GB, technically the former is faster because there is more bandwidth. The only time you will feel the difference, however, is when you are using integrated graphics for taxing purposes (ie: gaming). Because integrated graphics rely on the system's RAM, increased memory bandwidth will directly affect graphics performance.

 

Warranty is what matters most IMHO. Kingston has lifetime warranty. But just stick with name brand, Kingston, G.Skill, Corsair, Crucial, Samsung, Mushkin, PNY...

 

As a rule of the thumb, I stay away from "no name" brands as well as from "Value RAM" offerings from manufacturers such as Kingston.

The reasoning is twofold:

a) If one is building/configuring a system that is expected to last any reasonable amount of time (2-3 years), a $50-60 difference between choosing something well-known and properly tested/supported (Mushkin, Crucial, Samsung, Hynix...) and a "feebay special" is simply not worth it.

b) Vast majority of my customers use their laptops for serious work and bill their own clients at such rates. I can't have someone's machine go into BSOD in court because I wanted to save them $50 on initial build...

My $0.02 only...

 

I remember seeing a site where they did testing of various brands of ram, from name brands like Kingston and GSkill, all the way down to some no-name brand that I've never heard before in my life. They found there definitely was a performance difference between the name and no-name brands, but I can't remember how much the difference was. But I do remember the punchline being "unless the no-name brand costs half as much, it's not worth it to cheap out on the ram".

I'll try to dig up the review if I can...

 

Even if there isn't a performance difference, there's always questions of quality. While the RAM chips are pretty much all from the same sources, assembly of the modules are not.

 

From the first time I built a desktop to the present, I always just bought the cheapest RAM with the speeds and timings I wanted. Never had a problem.

 

Good lord yes, good branding will usually ensure that your RAM doesn't use cheapo, lower binned Memory ICs.
However, equally important also is the type of memory ICs used. Like it or not, all RAM manufacturers will have to use either Elpida, Hynix, Micron or Samsung (Nanya went OOB sometime ago). Each manufacturer has a manufacturing nodes of varying degrees of maturity and performance tuning priorities.

Crucial is quite reputable since they are a subsidiary distributor of Micron. Micron ICs are very reliable, they may not be the very best at power consumption or overclocking but they are very cost-effective. You definitely can't go wrong with Crucial.

Samsung ICs are arguably the most overclockable, power efficient and hence desirable due to their godly manufacturing capabilities. However, finding a manufacturer that exclusively uses Samsung ICs are a total crapshoot. My Clevo dealer, Metabox, almost exclusively used OEM Samsung RAM in all their laptops in the past but sadly, it seems of late the supply has dried up.

Hynix ICs are probably the next go-to in terms of overclocking. Hynix modules are arguably as good as Samsung but its rather difficult to find them. They overclock well but are just a little bit short of the sheer consistency of the Samsung ICs. Hynix memory ICs are preferable in performance oriented RAM modules.

Elpida ICs are basically Micron but binned for even greater cost-efficiency. They typically only work at their rated speeds and thats it, period, they simply overclock like crap most of the time. However, they are cheaper and you'll tend to find them on OEM RAM sticks or budget modules.

The above being said, the only real hard and fast rule is try to avoid modules that are too cheap. The risk of reliability issues having to use poorly binned ICs outweigh the extra $5 or so required to get better ICs.

Seriously, the performance difference between 8Gb and 4Gb modules are minuscule. You will notice the performance hit from lacking RAM far earlier than you will notice the performance hit from a few extra nanoseconds of latency.

 

What Wingnut said.

You could also find more esoteric reasons, too. I recently discovered that Corsair has an office fives minutes from where I live, so when I had to RMA something I just stopped by to drop off the package. Shipping's only, what, five bucks, but five bucks saved is five bucks saved. Nowadays I tend to look at Corsair offerings first just because of that.

 

For my customers i look to get longest warranty and easyest way to return. That 1-2% extra bandwith you will get dosent matter when you dont have usable pc for a week.

In my country kingston and kingmax offer 10year warranry so if you decide to upgrade or sell the system after few years you will sell those faster than various gskill/corsair that offer only 3 years.

Overclocking is something that 95% of users dont even care about.

Sent from my C1905 using Tapatalk

 

It's quite simple, like any other computer hardware is manufactured, you get what you pay for. "Brand" name typically buy up the highest binned product while lesser brands buy up lower binned products and can sell them at much lower cost. You also have to factor in customer service/warranty, lesser brands tend to have much worse customer service vs brand names. It's pay to play.

 

I ran into a sneaky problem that's worth knowing about --
I ran MEMTEST.EXE; it's commonly on Linux boot menus.
To my astonishment, MEMTEST began seeing memory errors. Most of it was sticky bits.
This surprised me so much because while I endured the 80's and 90's, memory had its flaky moments, but I thought this had been overcome.
The actual problem turned out to be this: MEMTEST hammers on the D-RAM chips so hard they get hot (really hot) and begin failing from the heat.
So MEMTEST *caused* indications of memory faults.

See why that's so blasted nasty?

-- thanks,
Dave Small

 

What about a expensive memory like the Kingston Hyper X with wrap speed! It has 1866 mhz. Would they give me extra fps compared to a memory from same brand but only 1600mhz?

 

Only if you were playing games on an integrated GPU.

 

Dedicated GPUs have their own RAM, so system RAM won't do anything that'll help gaming.

 

Except for games like BF4:

Battlefield 4 Loves High Speed Memory
Battlefield 4 - Memory Performance | bit-tech.net

But it's still marginal. Looks like Corsair over exaggerated the results where bit-tech shows the reality, lol.