OC Report :: OCZ Blade 3x2GB PC3-17000 CL8

[3oh6]

My last OC Report just over a month ago was on a kit of memory I referred to in a review as the Goliath of memory from the first testament. This OC Report isn't about some mere biblical fairy tale character, it is an all out Greek myth. Except this Greek hero isn't some weak PC3-16000 kit of memory - - this kit is straight from the top of mount Olympus. Hanging out with the Greek mythological gods, these sticks put Hermes and his winged shoes to shame. Clocked by Nomos on the 401 - like a chickenshyt cop in an airplane - doing over 1067MHz an hour on their way down to London-town, the Blade PC3-17000 8-9-8-24 kit has taken over the reigns as the deity of mount OCZ.


Let me simplify it for those too slow in the crowd to keep up, or those that simply slept through their first year Greek mythology class in university...they are rated for absolutely ridiculous clocks. The PC3-16000 CL7 OCZ Blades were great. They offered top bin specs and impressive overclocking head room...but they got nothing on these sticks. Well, maybe they do, but that is what I am going to find out today.

Right now is normally the point in an OC Report that I tell you what an OC Report is. That won't be the case this time. If you don't know by now, then you need to pull your anterior end out of your posterior end, take a deep breath, and have a look through my signature. The answer to the question "what is an OC Report?" might just be in there bookended by insults and rude gestures. With that out of the way, here is the table of contents for the afternoon...

Directory:

• Specs :: 36-24-36...That's What She Said on the Phone
• Photos :: No Centerfold But Close
• Test Bed :: It's Classified
• Go Time :: 32M SPi OC Report
• Nothing Stuffy Here :: Benchmarking Results
• Conclusion :: Amazingly Shallow Thoughts...By 3oh6

To finish off the full disclosure statements, or start them I guess, my agent tells me I have to tell you all that this kit was supplied to me by OCZ Technology Inc. for review purposes. That's right, I am such a big time professional overclocker and benchmarker that I have an agent. With recent talk of an F1 overclocking league and all these losers whining about HWBot becoming too hard to compete in I figured I should go out and get an agent. His name is Drew Rosenshack and he has advised me to simply say "no comment". If you don't get, quit being a geek, get some sun, and watch Sports Center once in a while

[3oh6]

Despite the big hype and the overcompensatingly large heat sinks that sometimes accompany modules these days, memory has the smallest spec sheet in the industry. Usually the entire spec sheet can fit within the tiny confines of a jail cell sized sticker on the module. Because of the "EnJoY" size spec sheet, this section is pretty small. Also, like EnJoY, I will be using a pair of balled up socks to enhance the appearance of the specifications section. Here first is a list of specifications plucked fresh from the OCZ web site.


Ohhhh, ahhhhh, look at those specs. DDR3-2133 means a ridiculous stock frequency of 1067MHz. Forget overclocking, just getting this memory to run stable at stock clocks is going to be a futile battle for many. I appreciate OCZ for putting this amazing kit out there, but someone clearly didn't communicate with the technical support division because they wouldn't touch these sticks with a ten foot pole. Good luck dealing with average Joe who can't wrap is mullet around the fact that his processor is the reason he can't run the memory at 1067MHz. I will be keeping my eyes on the OCZ technical support forum and relay any amusing threads that may pop up because of our new best friend here, the OCZ Blade PC3-17000, also known as OCZ Blade PC3-techsupportnightmare.


Oh, and by the way, if you haven't guessed, there are some highly binned Elpida Hyper IC's under the hood of these bad boys. These aren't your grandma's Hyper either. These have the individual throttle bodies, dual Garrett T3/T4 turbo's, lumpy cams and a 3" straight through exhaust. These Elpida Hyper's were binned for these sticks and these sticks only. This obviously means I now have at my disposal some of the best DDR3 memory that money can buy...and boy will you need to spend a lot of money if you want them too.

With crazy specs, and thorough binning, comes equally thorough prices. Here is what I could find for pricing at the time of posting...

• Sohodiffusion.com - $433.90 CND
• Mindfactory.de - 348,90 €

Obviously availability is limited with nowhere in the US showing stock through Google search. A few European sites are listing this kit at 348€+ and only Sohodiffusion.com listing in North America right now, at not a 'bad' price. Say what you want about the prices but if you don't like it, don't buy them. Try to walk into a Beamer dealership and walk out with an M3 for the same price of a Volkswagen Rabbit. As the saying goes, you gotta pay if you want to play.

[3oh6]

I will do what I can with the photos to bring something new, but there is only so much a lens can capture when taking the same photos. Actually, there is something new that wasn't in my last OCZ Blade OC Report which you can find at the end of this section. If you read the OCZ Blade PC3-16000 OC Report just scroll to the bottom for the new content, and if you didn't read it, I suggest you get over there right now and do so...I'll wait.

click for full size...

If you are reading this then you clearly can't follow directions. You really are the bottom of the evolutionary barrel sort of speak. Because of this, I will keep my comments simple, so you can understand them. Memory come in box. Box have logo on it. Inside of box is memory. Memory protected by plastic. Plastic is shiny. You like shiny...don't you? Yeah, shiny fun.

click for full size...

For those seeing the Blade heat sinks for the first time you might be a little under-whelmed, I know I was when I initially saw photos of the modules. The reason being is that the photos don't do these heat sinks justice. The finish on these modules is almost granite like and the flat black color really gives them a rugged feel. Aesthetics are obviously a personal thing and some people will like them, these are the cool people with a good sense of design. Others will think they aren't "fancy enough", these are the un-cool people with LED's and UV lights in their rigs...you know who you are. You think it's cool but really it isn't. You want cool, go to MurderBox.com and see what cool design really is. You stuffing a couple UV lights in your crap Ultra Aluminus with no cable management is a hack job...like Jimmy Fallon is to late night talk shows. Stop posting photos of your "mod", stop it right now...I don't want to see that crap posted in a forum ever again.

click for full size...

Forget about the first photo - you have all seen it before - but look at the second one. If you noticed the bottom of the heat sink pulling away from the module slightly...you would be right, congratulations, you win an imaginary cookie. Only one module is like this but it kind of sucks to see. The back side of the heat sink isn't actually flat and looks as if someone tried pulling the heat sink off. Needless to say, the bottom portion of the IC's are not contacting the heat sink. As if heat sinks on memory actually did anything, but if they do, then these modules heat sinks wouldn't be quite as effective as it should be. I just popped the back side off, flattened it out with a bit of force and all is well again. While I had the back off I snapped a quick photo of the modules...mmmm, Elpida Hyper.

[3oh6]

Like the last OCZ Blade OC Report, this one will also be done on my trusty EVGA X58 3X SLI Classified. I pulled the DFI X58-T3eH8 out of its box again for the OC Report and Review of this memory but it still hates me. Either I suck or it really is pissed that I ditched it for the Classified. My money is on the fact that I suck. Either way, I can't give that board away so I thought I would give it another chance, I am up to about chance four or five now, but it is still being a bytch about it so back in the box it goes. It is ugly as sin anyway so it doesn't deserve to be graced by the presence of these hot OCZ Blade modules. Here are some photos of the setup...

click for full size...

I have switched over to my recently won Intel Xeon W3540 processor for this OC Report as I wanted to see if the 920 was holding back the PC3-16000 sticks.


As you can tell, I am using a single stage on the CPU because the Classified isn't allowed to run with just air cooling anymore. A couple weeks ago I sat down with the Classified and it said that it didn't think it was being utilized properly with just air cooling on the CPU. It figured without cold, it just wasn't worth running. Who the hell am I to argue with a motherboard that is talking to me, right? Needless to say I have scheduled a therapy session for myself because, let's face it, I am having conversations with a motherboard; something isn't right. Here is a complete list of hardware used for this OC Report:

Hardware used:
MB:          EVGA X58 3X SLI Classified (BIOS S21S)
CPU:         Intel i7 920 (3845B026)
CPU Cooling: Chilly 1 single stage
GPU:         ATI Mach64 PCI
NB Cooling:  Stock
SB Cooling:  Stock
PWM Cooling: Stock
PSU:         Corsair HX1000W (single EPS connector used)
HD:          Seagate SATAII 80GB 8MB NCQ
OS:          Windows XP Pro SP3 - hopped up on 32M tweaks

Ambient Temperature: 22-23C

With that said, I think it is time to see what these sticks will do. Testing at 6-7-6, 7-7-6, and 7-8-7 is on the menu again with the added bonus of 8-9-8 this time as well. Time to see whether these sticks hold up to their top bin title, or if the PC3-16000 Blades show them up.

[3oh6]

I am honest about my testing, I am honest about my results, I am not trying to pull any kind of wool over anyone's eyes, unless of course she is hot and wearing a wool sweater. My OC Report simply shows one level of stability, single 32M SuperPi. I am not trying to pretend single 32M SPi stable is Prime Blend or LinX stable...it is what it is. This is not a review with actual stability testing, this is a report on overclocking, hence the name OC Report. I choose 32M SPi as my gauge because that is all I honestly care about. Complain, ask, whine, or mumble whatever you want about my testing...I don't care. Again, unless you are a hot female. Then we can discuss your concerns in private over dinner or something.

For the last time, these modules were provided by OCZ for review purpose. As I said last time..."Please refer to section one, paragraph four, line three, word thirteen for further details on what you can do should you feel this kit is cherry picked". OCZ has no agenda in sending out cherry picked kits of memory, I get random retail samples, nothing more and hopefully nothing less.


6-7-6-X 1T:
Intel X58 - EVGA Classified:
Click link for full size screenshot...
897MHz :: 1.657v | 922MHz :: 1.701v | 942MHz :: 1.758v | 962MHz :: 1.802v | 987MHz :: 1.856v

Ranging from 897MHz at 1.65v up to 987MHz at 1.85v, this kit is almost a dead ringer for what the PC3-16000 OCZ Blade sticks did. I looked at those almost a month ago and at this point, you could almost accuse me of doing nothing but laying new labels down. This time I didn't forget about my RTL and worked these sticks a little more. As I said about the PC3-16000 sticks, we are off to a great start.


7-7-6-X 1T:
Intel X58 - EVGA Classified:
Click link for full size screenshot...
1029MHz :: 1.657v | 1047MHz :: 1.701v | 1059MHz :: 1.758v | 1071MHz :: 1.802v

Opening up CAS Latency to 7, we can see the immediate jump in clocks. The quick standard test for Elpida Hyper to know whether they are going to be a solid universal kit is to see what they do at 1.65v and 7-7-6 timings. If you crack 1000MHz, you got a decent kit, if you crack 1020MHz, you got a good kit. These Blade PC3-17000 modules almost hit 1030Mhz at 1.65v so they definitely look like they are going to hold their own as the timings get looser. Unfortunately like the last OCZ Blade modules, these ones also died out at 1.75v~1.80v. No amount of wrangling with sub-timings allowed me to get them to scale up to 1.85v, but with that said, they managed to climb up to 1071MHz which is a solid 10 clocks higher than the PC3-16000 Blades did.


7-8-7-X 1T:
Intel X58 - EVGA Classified:
Click link for full size screenshot...
1035MHz :: 1.657v | 1065MHz :: 1.701v | 1089MHz :: 1.758v | 1113MHz :: 1.802v | 1131MHz :: 1.856v

I am not trying to sound like a broken record but my whole goal with this kit was to compare them to the results the PC3-16000's put up in my last OC Report. Soooo, like those PC3-16000 Blade's, these little bastards start out at almost the same clocks as 7-7-6 but as the voltage increased, they continue to scale at 7-8-7. Topping out at 1131Mhz with 1.85v is pretty damn nice but I was hoping for a little bit more. It just goes to show that the best of the best Elpida Hyper is pretty damn close to the best of the rest of Elpida Hyper. Perhaps I have been spoiled by two killer kits so my excitement seems lacking but that is about to change as I talk about the 8-9-8 results.


8-9-8-X 1T:
Intel X58 - EVGA Classified:
Click link for full size screenshot...
1161MHz :: 1.657v | 1179MHz :: 1.701v | 1197MHz :: 1.758v

Hahaha, what a joke, these modules won't even do single 32M at 1200MHz with 1.75v . <- that winky face means I am joking, so just relax all of you without a sense of humor. These sticks just love 8-9-8, even more than I love pretty ladies. I started at 1100MHz at 8-9-8 and thought something was wrong when I got to 1150Mhz still at 1.65v. Needless to say everything checked out and these sticks just absolutely destroy the IMC on this Intel Xeon W3540. The W3540 has a better memory controller than my i7 920 used in my previous OC Report and it still couldn't keep up with these sticks. To go any further I am going to have to push VTT up to 1.60v+ and I won't be doing that for a bit yet as I want to make sure I get something out of this chip before I kill it.

[3oh6]

Don't worry, there won't be any boring review like benchmark results showing how little of impact memory actually has on performance and what a waste of money high performance memory really is. All I will be posting here are the benching results that this kit is a part of as time goes by. Consider this the personal hall of fame for this particular kit of OCZ Blade 3x2GB PC3-17000 8-9-8-24 memory.

click for full size...

[3oh6]

Another day, another killer kit of Elpida Hyper from OCZ. As mentioned a couple times, this is my second kit in a row of OCZ Blade memory sporting Elpida Hyper IC's. The PC3-16000 Blade's got me smiling a month ago and these sticks have lengthened that smile to an all out ear to ear shyt eating grin. The results are on par or a little better than the PC3-16000 sticks were and really show that either of these kits are ready for the big time.


Unfortunately it doesn't seem to be only light at the end of the tunnel. As everyone is probably well aware of by now, a number of manufacturers have renounced the church of Elpida Hyper and ceased production of modules bearing these IC's due to premature failure in the field. Some manufacturers seem to be affected more than others, perhaps in part due to their screening, but even those that aren't seeing above average failures are cautiously moving forward without Elpida Hyper going to market. OCZ is one of those companies that doesn't seem to be affected, but in their official statement here, they do say they will cease Elpida Hyper modules for the time being. This puts the shelf life of these OCZ Blade PC3-17000 modules in jeopardy. So if you want them, get them now...or forever hold your piece.

[Bei Fei]

Nice run with the 2133 sticks I will try to push my sticks higher and wee what I can get from them!

[[XC] gomeler]

Post count whore

Nice piece of kit but in the end I'm not so sure about that price tag compared to other Hyper kits. I guess it'll be cheap to pick up some 2nd hand modules when all the average joes list them on eBay

[[XC] Oj101]

Where the hell did they find you? That review had me in stitches, crazy drugs man!

[r1rhyder]

But like you said 3oh6, it all depends on the cpu's memory controller.

[3oh6]

Post count whore

that's me...the king of "nice clocks" or "well done" posts. unlike Mia thinks, post count does = intelligence up here in Canada...and i done needs me all the intelligamence i can gets.

Where the hell did they find you? That review had me in stitches, crazy drugs man!

ahhh Canuckistan, home of the beaver (yeah i said it, hot chicks everywhere boys), and home of BC...where they grow the best damn lettuce on the planet

glad you enjoyed the read

[Expat GriZ]

Jody......Hemingway has nothing on you!! Another great read and report. Seeing as I just had one of my Blades die on me, might as well try and exchange them for a set of these....now all I need is a nice 3540 or 70!!

[eva2000]

Very nice review Jody... just not sure how much of the memory head room is the modules and how much is aided by the sub zero cooling on the cpu ? Does sub zero cooling only help memory headroom frequency or also allow tighter timings at higher frequencies ?

Still crazy validation @DDR3-2440Mhz 8-9-8-x!

Shame the Elpida Hyper IC failure issues in some batches are creating doubts in peoples' minds. Still Elpida Hyper ICs are the only ones that can clock that way - unless the new Micron 2000Mhz binned ICs improve ?

[chispy]

Incredible speeds out of this modules , as always good job on your OC Report jody .


Angelo.

[Leeghoofd]

Job well done Jody, damn these things can go hard...

[zsamz_]

jody would be nice if you can test them in a rex

[3oh6]

Does sub zero cooling only help memory headroom frequency or also allow tighter timings at higher frequencies ?

aren't those the same thing...ultimately higher frequencies, and depending on the IMC, cold can help a lot. if you are asking if cold not only help high clocks but tighter timings at the same clocks, that i doubt...unless of course the IMC was limiting the tighter timings.

Incredible speeds out of this modules , as always good job on your OC Report jody .

Angelo.

Job well done Jody, damn these things can go hard...

thanks guys, definitely a fun kit to play with. i was thrashing them in the overclockaholics.com 01 low clock challenge too. 1100MHz 7-8-7 all day long in 01 at 1.80v. could have tried lower volts but didn't bother.

jody would be nice if you can test them in a rex

i am a little tied up for time and i'll be honest, i haven't turned the RE since probably november so i am not sure i would ever get to testing on that platform...besides, i7 is better

[3oh6]

i finally had time to do some benching with this kit and achieved new 32M & AM3 personal bests from this kit at 1172MHz and an 1182MHz plus a 1M thrown in for good measure

click for full size...

[eva2000]

congrats Jody! awesome memory clocks DDR3-2344-2364Mhz!

[CryptiK]

Really nice work there. Awesome clocks and congrats on the PB's!

I would also be interested to see what they can do on a REX, but if you are too busy checking out the beavers and playing with your 'i7' then that's cool

In all seriousness though, the REX would just hold them back, I can't get 1000+ MHz stable with those voltages.

Would you really consider 1.80v 'thrashing' these Elpidas?

[zalbard]

Great clocks, and love your writing style!
Awesome review!

[gandalfone]

Great review Jody

[3oh6]

thanks guys, and those stuck on the RE...move on already, you guys are sounding like Gautam

[WaterFlex]

Excellent sticks! Good job! OCZ i s back!