[MemoryReaction] CellShock 2x1GB PC2-8000C4

[EnJoY]

Introduction:

Back in December of 2006, I contacted Alonso of CellShock for a review sample of their flagship product, the PC2-8000C4 2GB Kit. It took me until late in January to obtain my sample as they only had a limited number of samples they could give out, and I had missed the mark for the previous month. Lucky for me, once January came around I had already put my name in the queue and I was good to go. Regarding the sample I had received, I've been assured by Alex (Alonso) that my review set is a 100% retail representative sample.

I began testing the modules and I was very pleased with the results. However, half way through my testing, I was running Orthos at 4-4-3-9-3 with 2.4v and suddenly the system blue screened and wouldn't let me boot back up. After some trouble shooting I found that one of the sticks had died during the testing. I contacted Alex about the problem and suggested that it might be voltage that was the cause. However, Alex informed me that CellShock performs extreme heat and voltage tests on their modules for long durations of time to eliminate sticks that may be prone to failure from such things. Therefore we concluded that it must have just been one of those times when "it just happened", without much explanation.

Luckily for me, with a dead module, I was able to RMA my sticks within the states thanks to the very gracious Ben, of Tankguys.com. Alex and Ben worked out the details and allowed me to get a new set from Ben's stock to replace my defective set in short time. When my new set arrived, I tested each stick individually and found that the worse stick of the two in this set performed exactly the same as the worse stick of the previous defective set. Which of course meant that I didn't have to waste time repeating my previous testing. This also confirmed that the set I had received from CellShock directly was indeed NOT a cherry picked review sample, but in fact a retail representative sample just as Alex had stated. Thanks again to Alex and CellShock for their honesty and integrity in this matter as it is of the utmost importance to the readership of this forum.

Below here is a small blurb about CellShock as a company themselves taken directly from their website:

Originally Posted by CellShock Official Website

MSC GmbH was set up in Stutensee, Germany in 1982 specialising in the marketing of semiconductor products and developing new electronic subassemblies. With our numerous branch offices, we have an international profile to serve industrial customers in IT, power engineering, aviation and medicine with components from high-profile manufacturers as well as internally developed and manufactured products. We have also put major investments in research and production in recent years to apply our solid expertise to next-generation products.

Producing user memories in all variations and specifications for industry has always been one of MSC's core competencies. And now, with the new CellShock series, we are applying the knowledge gained to DRAM memory systems for PC fans. We use special techniques for chip selection and printed circuit boards with enhanced layouts including exhaustive quality testing. We also have contracts with a variety of reputable manufacturers of memory chips and electronic components that enable us to access a wide assortment of applications for optimising our products.

This calls for major product performance quality because the test programs we use were developed by overclockers for overclockers. We give that extra performance and compatibility - not because of some advertising slogans, but because we have a genuine passion for high-end PC components. We want to make a believer out of our customers not only with our products, but also fast service and expert technical support.

Now, onto the exciting stuff...the review...


The Memory:






-------------------------------------------------------------


Specifications:

Specification: DDR2 1000 (PC8000)

Timings: 4-4-4-12

VDimm: 2,1V - 2,3V

Chip: Micron D9GKX 64Mx8

PCB: 8-layer, 240pin

Guarantee: 5 years

-------------------------------------------------------------


Test Setup:

Intel Core 2 Duo "Allendale" E6400 - Lapped
Asus P5B-Deluxe rev1.03G - Stock, Bios: 910 Official
CellShock 2x1GB PC2-8000C4 "CS2221440"
BFG GeForce 7950GT 512MB @ 700/800
OCZ PowerStream 600w

Cooling:

CoolTechnica AQX MP-05 SP LE, Swiftech MCW60, Black Ice Pro III, Swiftech MCP655, 7/16" ID
120mm Yate Loon @ 12v over memory

-------------------------------------------------------------


Testing Methodology:

The testing process I used is quite simple, and I'm guessing quite close to the same procedures the bulk of you use to test your new hardware. I simply went into bios, and set the tightest timings I felt would be stable at the stock rated voltage, while still leaving me some head room to clock upwards. I then proceeded to exit bios and enter Memtest v1.70. I tested using tests 1, 2, 4, and then 10 loops of test 5 before moving on to the next speed bump. I worked in the 2:3 ratio and moved up 2MHz on the FSB at a time. I also used full vFSB and vMCH voltages the entire time to ensure the chipset wouldn't hold me back. I increased clock speed until I would error in memtest, then I'd back down 1MHz on the FSB at a time until the errors disapeared, then I'd log my results and push up the voltage another notch and do the same testing with that. Once this was done, I entered windows and began testing in Orthos using 2hr test periods (at a priority of 3) to attempt to document some level of stability; atleast higher than SuperPi 32M, while still maintaining my social life and other responsibilities. This time around my test results in MemTest proved to remain very true to those I received in Orthos. Meaning that the limits I was finding in MemTest were nearly identical to the limits I hit in Orthos testing.


The Review:

Each piece is separated by timings and voltages. I started with 3-3-3-6-3 and ended with 5-5-5-15-5 to achieve maximum MHz. I've included screenshots of MemSet this time with the results to show the sub-timings I chose for each run.

The voltages listed are what was selected in bios. These are not "actual" voltages.

-------------------------------------------------------------


Timings: 3-3-3-6-3

2.25v


2.35v


2.45v




Timings: 4-4-3-9-3

2.1v


2.25v


2.35v


2.45v

Could not get any clock speed increases with this voltage, even with adjustments to the sub-timings. This still confuses me...


Timings: 4-4-4-12-4

2.1v


2.25v


2.35v


2.45v



Timings: 5-5-5-15-5

2.1v


2.25v

No increase, chipset limited.

-------------------------------------------------------------


Conclusion & Verdict:

Well these sticks were certainly a blast to play with! Look at some of these results, DDR860 3-3-3! DDR1140 with 4-4-4 timings with 2.4v! And most impressive, DDR1160 5-5-5 with only 2.1v!!! And all on an unmodded P5B! I'm just very impressed with these sticks and the results they put out. Unfortunately I couldn't reach my goal of DDR1200s, but it was not the fault of the memory as can be assumed from the outstanding 5-5-5 timing results. No matter how loose I made the sub-timings, I was simply hitting a wall past 1165 or so that was begging me for more vFSB that I did not have.

I must say that overall these sticks are a jack of all trades. They can do 3-3-3 timings at stock voltages with room to clock higher, and if you want to go for all out bandwidth, they'll simply fly at 5-5-5 timings and above! I've found the sweet spot for these sticks for every day use and reliability is DDR1070 with 4-4-4-10 timings and 2.25v. However, any route you choose, you really can't go wrong.

Soon I'll be testing on an Asus Commando, so these problems with chipset limitations will hopefully be out of the way. And in a week or so, I'll be posting up some results with these sticks as well as the G.Skill HK's I reviewed earlier on the eVGA 680i board to see what both sets are capable of for high MHz!

If you'd like to purchase the CellShock 2x1GB PC2-8000C4 "CS2221440", you can get them from Ben at Tankguys.com. At ~$419.99 for the 2GB kit, I find these are a good value when compared against other top performing sticks such as Corsair's Dominator series or OCZ's Flex series.

I'd like to thank CellShock for this review opportunity. And of course, this has been a hardware reaction from yours truly, EnJoY.

[guess2098]

newer micron has problem to run 4-4-"3"

[s7e9h3n]

Originally Posted by guess2098

newer micron has problem to run 4-4-"3"

Care to explain Denny?

And @ OP....So basically you tested these sticks for stability for 2 hours with ORTHOS @ priority 3? Any spi results? Or maybe some memtest? Also, why did you choose to change only those subtimings and not any others? The memory is probably capable of a lot more....let's see some Xtreme performance!

[fhpchris]

Agreed Stephen

What are your subtimings?

[EnJoY]

Originally Posted by guess2098

newer micron has problem to run 4-4-"3"

Really? This is the first I've heard of this, how do you know this Denny?

Originally Posted by s7e9h3n

And @ OP....So basically you tested these sticks for stability for 2 hours with ORTHOS @ priority 3? Any spi results? Or maybe some memtest? Also, why did you choose to change only those subtimings and not any others? The memory is probably capable of a lot more....let's see some Xtreme performance!

Orthos for 2hours at priority 3 has shown to me to be 100% 24/7 stable nearly 100% of the time. It's also a realistic test time that allows me to complete the full review in a timely manner. Super pi is pretty useless to test as it really doesn't say anything for stability, except that it's not gonna crash on you in the next few hours, but it might the second you open a game or something. I can game at all of these settings without a problem. I've thought about adding 3DMark to the top results in each timing category to show the performance advantages/disadvantages of each, but I didn't with this as I want to determine what the best way to do that is.

I played with nearly every subtiming, but only kept the ones I saw benefits from. Trust me, I tried everything to try and break 1170 and I just couldn't. Dropping down to a 1:2 divider helped a bit, but still didn't stop crashes. It's just a chipset limitation at this point.

[s7e9h3n]

Originally Posted by EnJoY

Really? This is the first I've heard of this, how do you know this Denny?



Orthos for 2hours at priority 3 has shown to me to be 100% 24/7 stable nearly 100% of the time. It's also a realistic test time that allows me to complete the full review in a timely manner. Super pi is pretty useless to test as it really doesn't say anything for stability, except that it's not gonna crash on you in the next few hours, but it might the second you open a game or something. I can game at all of these settings without a problem. I've thought about adding 3DMark to the top results in each timing category to show the performance advantages/disadvantages of each, but I didn't with this as I want to determine what the best way to do that is.

I played with nearly every subtiming, but only kept the ones I saw benefits from. Trust me, I tried everything to try and break 1170 and I just couldn't. Dropping down to a 1:2 divider helped a bit, but still didn't stop crashes. It's just a chipset limitation at this point.

Trust me, Denny knows his $hit

2 hourse of Orthos @ priority 3 may be an indication of 100% stability for YOU, but that doesn't mean it'll be stable for others who may run more memory intensive programs. If you're gonna test using Orthos, you could at least try to use a more stressful setting as Priority 3 is pretty hard to guage for most here. I'm sure that I'm not the only one who's never even tried to test for stability using that low of a priority.

I never said that you should use Spi as a test for stability, but since I haven't yet seen threads about "WR for LgFFT Orthos @ Priority 3" yet, I'm sure more people would be interested in those results than most other tests out there

And correct me if I'm wrong, but don't the Asus Commando and the P5b use the SAME chipset? 965? And why not test on the 680i?

[EnJoY]

Originally Posted by s7e9h3n

Trust me, Denny knows his $hit

2 hourse of Orthos @ priority 3 may be an indication of 100% stability for YOU, but that doesn't mean it'll be stable for others who may run more memory intensive programs. If you're gonna test using Orthos, you could at least try to use a more stressful setting as Priority 3 is pretty hard to guage for most here. I'm sure that I'm not the only one who's never even tried to test for stability using that low of a priority.

I never said that you should use Spi as a test for stability, but since I haven't yet seen threads about "WR for LgFFT Orthos @ Priority 3" yet, I'm sure more people would be interested in those results than most other tests out there

And correct me if I'm wrong, but don't the Asus Commando and the P5b use the SAME chipset? 965? And why not test on the 680i?

I know he does, I just spoke to him on MSN on his comment on the Micron. Apparently he's seen the same wall I found in his own testing.

I can't crash anything after this testing MOST of the time. I'm not saying it's always 100% stable, but it's a good balance. It's takes a little longer than 32m, but means a lot more, and as I said before, testing with priority allows me to still use this machine (which I do) while it's testing. This is my only machine for the time being ya know.

And I understand the super pi thing, I just don't value it as much as you do I suppose. If enough people demand pi tests in addition or in replacement of my current testing methods, I'll give in. But people really seemed to like the way I did things last time.

And yes stephen, you are correct in that the Commando uses the same chipset as my P5B. But when I say chipset limitation, I mean mostly in regards to available voltages. I do not volt mod, nor do I have the hands for it. I'd have someone else do it, but the Commando has everything I need right out of the box, so why bother? I was asking Tony what he thought would be the best choice for testing memory and he agreed with me on the Commando being one of the best choices for this purpose.

And I will be testing a 680i, incase you missed where I mentioned that in my conclusion. However, I have to adjust my forcast as I won't be around at all next week (spring break cruise).

[guess2098]

Originally Posted by s7e9h3n

Care to explain Denny?

And @ OP....So basically you tested these sticks for stability for 2 hours with ORTHOS @ priority 3? Any spi results? Or maybe some memtest? Also, why did you choose to change only those subtimings and not any others? The memory is probably capable of a lot more....let's see some Xtreme performance!

just did some research of micron lately and found this super weird ..not really a problem but problem lol.

also Alex, P9 is actually better then P3 but yeah dual 32M for dual core is really good to tell it is stable or not.

by the way, isn't subtiming show on the memset?
it is 3-3-3-8-?-35-5-10-8-10 on his 1st post

5-10-8-10 isn't bad at all (average)

[s7e9h3n]

Originally Posted by guess2098

just did some research of micron lately and found this super weird ..not really a problem but problem lol.

also Alex, P9 is actually better then P3 but yeah dual 32M for dual core is really good to tell it is stable or not.

by the way, isn't subtiming show on the memset?
it is 3-3-3-8-?-35-5-10-8-10 on his 1st post

5-10-8-10 isn't bad at all (average)

Yep, I know he's SHOWING the subtimings, but I was wondering why he chose to change ONLY the 4 and not adjust any of the others

5-10-8-10 ain't bad, but for $400+, I'd expect to see something like this

[Quentin]

580 Cas 5 don't pass with lower vdimm ?

My Cellshock PC2-8000 Cas 4 :
Mobo : Biostar Tforce P965 Deluxe
Vdimm : 1.8v
Timmings : 5-5-5-12



I can't test with higher vdimm, chipset limited

[Blknoel]

Quentin: You use 1M Pi to conclude it "passes" a Memory Test?

At least dual 32M at 1.8v and see if it survives.

Btw, FSB is at 380 and you are chipset limited? :P (or you mean Bios options limited?)

[boblemagnifique]

My reviews : http://www.overclocking-masters.com/....php?id=52&p=3

6 kits of which (Cell Shock 2x1go pc8000 c4)

Protocol of test : Spi 8mo and 3d Mark 2001 under EVGA NF680i

No the max Stable(too much time) but that gives an indication

Great Mémory et Spécial Thk at Alex

[boblemagnifique]

Originally Posted by Blknoel

Quentin: You use 1M Pi to conclude it "passes" a Memory Test?

At least dual 32M at 1.8v and see if it survives.

Btw, FSB is at 380 and you are chipset limited? :P (or you mean Bios options limited?)

For Test a Stap 1066

That would be interesting to test out of strap 1333 but the ratio doesnt allow it under i965 (1:1 only)

[Quentin]

Originally Posted by Blknoel

Quentin: You use 1M Pi to conclude it "passes" a Memory Test?

At least dual 32M at 1.8v and see if it survives.

Btw, FSB is at 380 and you are chipset limited? :P (or you mean Bios options limited?)

Of course not

32M at 1.8v = 560-565.

I'm not limited by the FSB, the chispet don't want go over 570-575 for the ram speed with 1Gb stick. Same probleme with my OCZ PC8500.

With 512Mo stick : 625-630

I need P5B Deluxe

[EnJoY]

Originally Posted by s7e9h3n

Yep, I know he's SHOWING the subtimings, but I was wondering why he chose to change ONLY the 4 and not adjust any of the others

5-10-8-10 ain't bad, but for $400+, I'd expect to see something like this

Stephen...please, that's a 680i... The TRCD and TRP don't even effect stability in the same way on that board as they do on every other board. 4-4-4 or 4-3-3 has no stability impact. Also, every timing change on that board has a different stability and performance impact than it does on my board. And lastly, you never listed the vdimm used.

If you're gonna be ridiculously picky, I can humor you, not don't just be ridiculous.

[loc.o]

Nice tests Both me gskill and cellshock wont do over 571 4-4-4 on me unmodded P5B... I think you should be able to get a bit further with 5-5-5 though. vFSB seems to help me.

Didnt have that much time to test but this is what mine do with 2.45v. Wil give Orthos a go soon.

3-3-3 booted at 401 so 32M is slooooow


4-4-4


5-5-5

[s7e9h3n]

Originally Posted by EnJoY

Stephen...please, that's a 680i... The TRCD and TRP don't even effect stability in the same way on that board as they do on every other board. 4-4-4 or 4-3-3 has no stability impact. Also, every timing change on that board has a different stability and performance impact than it does on my board. And lastly, you never listed the vdimm used.

If you're gonna be ridiculously picky, I can humor you, not don't just be ridiculous.

Oh really?? So according to you, 4-4-4 would be just as easy to run as 4-3-3, right? And when I said "4" in my prior statement, I meant that you chose to change only those FOUR particular values. Now that I look at it again, you only adjusted those timings a few times and let the "Auto" function of memset do the rest....

Hey, I'm not being rediculously picky, I just wanted to see just how much performance could be squeezed out of this set of RAM. This is a memory review - as you say - so it'd be nice to see just how hard this memory can be pushed, regardless of your "limitations". The memory manufacturer chose to send you a set of RAM in order to possibly publicize it so I'd think that they'd at least like to see a bit more effort in the review. Let me ask you this....what exactly did you do with your "review" that ANYBODY with a motherboard and a cpu COULDN'T have done themselves? If I wanted to see factory performance, I'd go to Cellshock's website or Tom's Hardware or Anandtech.....I come to Xtremesytems to see PERFORMANCE numbers, not reviews in which all somebody does is raise the Vmem to gain "stability"......Maybe next time you do a write-up, you really should take into consideration who your target audience is. I'm done here...Thanks for the lesson on timings...I'll keep that it mind - hopefully one day I'll be able to overclock memory well.

[erwinz]

nice review mate..

thanks for sharing.. I like orthos testing much more than dual 32m pi this is just me..

[EnJoY]

Originally Posted by erwinz

nice review mate..

thanks for sharing.. I like orthos testing much more than dual 32m pi this is just me..

Thanks erwinz, I spent a lot of time on this review, so I'm glad you could benefit from it.


As for s7e9h3n's comments. It really upsets me to think a legend, such as him whom I respect, can show such disrespect toward me, when I'm only trying to help the community in showing them an unbiased retail outlook on what they can expect if they had purchased these modules. The goal of my reviews are not to show what the sticks do with 2.7+ volts in a spi 1m benchmark...but to show they can do on the average users configuration under standard stability testing for every day use. If any of you feel I've failed in accomplishing this, please let me know and tell me how you'd prefer me to test in the future.

I'm not claiming in these reviews to be some tweaking god, that's s7e9h3n's forte. I'm just claiming that I squeezed everything I could out of these modules on my standard unmodded setup to the best of my knowledge and abilities.

[tzitzos]

always enjoying read your test enjoy
keep up the good work and thx for sharing

[Stelios]

Enjoy , does the "large FFT" stress rams more than blend ???

PS . I bought these rams 3 days ago and i'm in the "burn-in" process right now .

ΤIA .

[EnJoY]

Originally Posted by Stelios

Enjoy , does the "large FFT" stress rams more than blend ???

PS . I bought these rams 3 days ago and i'm in the "burn-in" process right now .

ΤIA .

It should. And from my own personal testing it's also been proven, which is why I use it.

[scarface]

Originally Posted by EnJoY

Orthos for 2hours at priority 3 has shown to me to be 100% 24/7 stable nearly 100% of the time. Super pi is pretty useless to test as it really doesn't say anything for stability...

So true. They'll bash you for using orthos for 2 hours then turn around and use weak a** superpi for 13-14 mins and call it stable.
Funny how spi only uses 50% of each core and i've rarely seen two instances of it being used. Whenever i see o/c's with Superpi showing, like the ones above, i don't even pay attention to them.
I've actually had my pc bsod after 2 1/2 Hours of orthos, so what does that say about a 14 minute or less load at 50%.

With that being said, i thought it was a nice review, Thanks EnjoY.

[erwinz]

Originally Posted by EnJoY

It should. And from my own personal testing it's also been proven, which is why I use it.

mm.. Ive been testing my rams lately..

I use memtest86 boot from cd.. do you think its much better test??

[EnJoY]

Originally Posted by erwinz

mm.. Ive been testing my rams lately..

I use memtest86 boot from cd.. do you think its much better test??

Memtest is what I always start with, and after I determine what is stable in Memtest, I move on to Orthos testing.