PNY MEMORY VS ANY OTHER

PNY is just another memory reseller. I've purchased their Compact Flash cards and had zero problems with them.

Their memory modules and/or chips may come from any number of different suppliers. Most seem to be relatively well known name brands and some reputable second-tier manufacturers. Some of their chips are marked with the PNY brand name, although it's unclear if this is generic "bin-sorted" memory or simply first-tier memory that's been house-branded.

I guess the thing with Crucial is that you know you're getting Micron memory most of the time, but at least another top-tier memory manufacturer like Samsung. They also "burn-in" test every module they sell, while some others may simply "sample test".

Kingston memory is almost always from one of the top-tier manufacturers like Micron, Samsung, Hynix, Mosel-Vitelic, or Infineon. They say they test every module and chip, but don't indicate if their Value RAM series is "burn-in" tested.

I'd have no problem with PNY memory if I got a good price on it. However - if Crucial's prices are comparable, I'll go with them every time.

PNY doesn't make any of their own memory. They buy their chips from Micron, Samsung, Mosel-Vitelic, Hynix, Infineon, Elpida, etc. Sometimes the manufacturer is identifiable, or sometimes PNY has their name printed on the chip packaging. I don't even know if they have their own board making facilities, although it wouldn't surprise me either way. Their function is to solder on the chips to the circuit boards, test them, and package them. They also verify that certain modules work in specific applications.

Crucial is really just a retail division of Micron. The Crucial module I recently bought had a Micron sticker on one side and a Crucial sticker on the other. All chips were Micron (MT logo). I just consider Crucial to be the "gold standard", although I would buy PNY if I could get a substantial discount.

Perhaps it's a matter of definition. I consider a "memory maker" to be a company making the silicon wafers and packaging them in plastic or ceramic parts. PNY is an assembler. They take parts and put them together.

Even PNY might buy fully assembled modules and resell them if they have a shortage or can get a good price on quality parts.

Hi y pw,

I thought you might find this interesting. AnandTech had a tour of Kingston Taiwan a few months ago. It appears they test each and every module by machine as well as by hand as well. The end boxed product was a ValueRAM module.

I've bought quite a bit of Kingston ValueRAM over the years and haven't been bit once.

Benny Li:

i I thought you might find this interesting. AnandTech had a tour of Kingston Taiwan a few months ago. It appears they test each and every module by machine as well as by hand as well. The end boxed product was a ValueRAM module.

Sure - testing modules isn't all that difficult although some of the bargain brands are simply sample tested and the cost savings on testing go into replacement of modules that fail early. Standard testing of all modules is the next step. Perhaps you already know this, but a true "burn-in" (or accelerated life) test is generally what provides the highest assurance that a part will be reliable in the long term.

Even Kingston has "burn-in" ovens although I don't believe they use them for the ValueRAM series. It's just a matter of the reliability "bathtub curve" one sees with most electronics. After weeding out the truly defective parts, the life before failure follows a curve where a few parts will fail early, the large majority will last reasonably long, and a few extraordinary parts will last even longer. By baking them in ovens to accelerate the aging slightly, the manufacturer can test and weed out most of the early failing parts. It costs more and perhaps reduces the ultimate life of the parts, but it vastly reduces the number of early failing parts that are sold.

http://www.tutorialsweb.com/reliability/reliability1,2,3.htm

$150 is expensive for a 1GB kit. Many reliable brands are selling for under $100. For example:

<http://www.newegg.com/Product/Product.asp?Item=N82E16820145440>

<http://www.zipzoomfly.com/jsp/ProductDetail.jsp?ProductCode=80098-20>

<http://www.chiefvalue.com/app/productdetails.asp?submit=property&item=20-236-11 0>

Thanks, Fred.

For us, the return policy at newegg, zipzoomfly, owc, etc. is fine because we don't order anything unless we know exactly what we want. We have never returned an item that was subject to restocking. The few times we have received a defective item we received a replacement promptly at no additional cost. DV Forge even shipped a replacement the same day I contacted them - before I even returned the item. This was before they acquired their present reputation, so we don't recommend them now.

In all the years we have shopped on the net, we have yet to have a bad experience. The main reason is that we do a lot of research before we buy from a site the first time. If you can get the exact item you want significantly cheaper from a reputable source, why not save the money? If you are not sure, you are probably better off to go to a local store.

Jo Ann

Hi Matt,

The enthusiast RAM offered by those companies are generally of lower latency and have higher rated chips; they also have heat spreaders to help cool off those higher rated chips and possibly the use of higher voltages. They're geared towards speeds beyond DDR400, speeds that have yet to be standardized.

The use of enthusiast RAM is mostly pointless on a Power Mac G5 as there is no easy way to fine tune installed RAM like with a PC BIOS. If you do want to get enthusiast RAM for the G5, look for DDR400 RAM that can run 2-2-2-x or 2-3-2-x at 2.5-2.6v as well as Apple SPD compatibility. This is especially important because the G5s check for CAS3 and will not start without that information on a stick of RAM. The RAM will run at the faster speed after the computer has finished booting, but CAS3 must be programmed on the stick.

As for reliability, the value RAM from most reputable RAM assemblers or manufacturers would be on par with the enthusiast stuff provided you don't push value RAM past its rated spec. 😉

FYI,

$20 rebate on PNY 2x512MB kit for G5 (Part# D1GBPC32-G5KIT) till Oct 14:

http://www.newegg.com/Product/Product.asp?Item=N82E16820236110

$69.44 after rebate (I am buying two for a total of 4x512MB)

I did a lot of research, and for the G5, it seems that the consensus is that it cannot use a "faster" i.e. CL2 memory, and the recommendation is to buy memory strictly by price so long as the manufacturer guarantees compatibility with the G5. Not all PC3200 might be compatible with the G5, strict adherence to JEDEC is required, and the SPD has to have CL3 in it.

http://www.anandtech.com/mac/showdoc.aspx?i=2359&p=1

good article, a bit old (esp price wise) here is a short quote from it:

"Although all current generation Macs use PC compatible DDR memory, Apple adheres strictly to JEDEC specifications and requires very strict adherence to their timing specifications for module compatibility. Apple ensures that all memory modules meet their timing requirements by not POSTing if incompatible memory is installed in the system. So, it is very important that you only use memory that's specifically made for the Mac that you're trying to upgrade.

Because all Macs configure their memory timings based on the SPD that Apple specifies, all memory is configured to have the same timings regardless of what the modules are capable of - in other words, all Mac memory performs the same. It's not possible to overclock the memory bus on Mac platforms either, so the maximum frequency capabilities of the modules don't matter either. All memory vendors these days offer lifetime warranties on their memory, so what it truly boils down to is compatibility and price - which is exactly what this guide is designed to compare...."

though there is one conflicting report from http://www.barefeats.com/quick.html

"August 18th, 2005 -- Maybe memory latency speed makes a difference after all. I was doing some testing on two different G5/2.5GHz Power Macs. They are configured identical in every way (8GB RAM, GeForce 6800 Ultra). In fact, they have identical Raptor 10K boot drives in the top factory slot with contents of one cloned to the other. When I ran my Motion 2 "render RAM preview" test, one system ran 8% faster than the other. Why? Well, there is one difference. The faster G5 had CAS 2 memory and the other one had CAS 3. "

-avi

i "August 18th, 2005 -- Maybe memory latency speed makes a difference after all. I was doing some testing on two different G5/2.5GHz Power Macs. They are configured identical in every way (8GB RAM, GeForce 6800 Ultra). In fact, they have identical Raptor 10K boot drives in the top factory slot with contents of one cloned to the other. When I ran my Motion 2 "render RAM preview" test, one system ran 8% faster than the other. Why? Well, there is one difference. The faster G5 had CAS 2 memory and the other one had CAS 3. "

CAS latency could make a measurable (but not earthshattering) difference in truly memory intensive applications that primarily pull data from the system memory.

My guess is that the Power Mac G5 reads are the maximum DDR burst length of 8.

A CL2 would read 8 "words" in 2 .5+.5+.5+.5+.5+.5.5=5.5 clock cycles.

A CL2 would read 8 "words" in 3 .5+.5+.5+.5+.5+.5.5=6.5 clock cycles.

So the CL2 would read about 15% [(6.5-5.5)/6.5] faster. However - there should be no difference for write timing (1 .5+.5+.5+.5+.5+.5.5=4.5 clock cycles). This performance benefit probably isn't that noticeable in other applications.

I just bought a 1g kit form CompUSA for $150 that seems to work fine. I plan to buy more as $ allows.

Does anyone know how big a difference there is between PNY or Corsair Value memory and the premium memory from those brands. Is it mostly an issue of stability and reliability?

Thanks,

Matt