The brand new Corsair Platinum series continue on the legacy of the Corsair high end series. Instead of reusing the typical Dominator GTX red and black look, Corsair decided it was about time for a major esthetical overhaul, and boy these sticks sport a killer look. The black heatspreader design in combination with a LED illuminated aluminium chrome top bar make these new Platinums truely stand out from the crowd. The box design is also a big step forward from the typical blister or cardboard versions Corsair previously tended to use.

 

 

 

The 4 x 4Gb Dimms are well protected by it's flashy box. Each Dimm inside a small blister package to avoid scratches, dust collection and probably the most important: drooling fluids from the PC shoppers.

 

 

It has to be noted that these Platinums are NOT intended to be the new high end enthusiast series from Corsair; there's more to come soon at that level, seeing the daylight in probably a new GTX range (why oh why did we have to wait soo long ?). The Platinum series are designed to be the perfect harmony between looks, bling bling, performance and reliability. The Platinum line up looks like this:

 

 

High out of the box frequencies and all sorts of different volumes are adressed in the new Platinum lineup. Too bad there are no 8GB kits for the 2400 and 2666Mhz frequencies, as this would drastically reduce the cost of these kits. But 489 dollars for the high end kit 16GB 2666C10 version seems justified, if it performs adequately :)

Our 2666C10 kit is based on Samsung IC's according our PR contact. Looking at the achieved OC results we have to agree unanomously. For the the other kits we gladly link you to an adjacent forum as there are wild discussions on which dimm sports which IC's. This link at XS forum also shows some achieved results form daily tweakers and overclockers.

 

Due to the fact Corsair tends to glue their heatspreaders to the IC's no pictures were made of the bare Dimms. However the chrome top integrated light bar is swappable. Allowing different optional LED colours, to really top off your case mod. The optional light bars should be available somewhere in January 2013. Secondly these Dimms can also be hooked up via the LINK software. Allowing the user to read out the IC activity and temperatures. Not really of any benefit to me, but some people like to be in control of everything :)

 

 

Before we go to the testing some more pictures of probably the most sexy Dimms around on our blue planet.

 

 

 

 

 

 

 

While it is always hard to recommend or even test a high end kit, we tend to include the results of at least a mainstream kit. In this case 1600Mhz C9-9-9-27 2T Command Rate was used. The results included in the charts are simulated Platinum versions, replicated by this Samsung based 2666C10 kit. Stress is on the word simulated, as lower end versions are mostly Hynix and Elpida based, so results can slightly vary. The purpose is to give you an idea of what more RAM speed brings to your rig.

Our test setup comprises of the splendid ASUS Maximus Z Gene V board. One of the best boards around with very few quirks. The RAM compibility and tweaking capabilities are also second to none. From all the Z77 boards I have had the honor to test, the ROG Z77 versions are pretty untouchable at the moment.

• ASUS ROG Z77 GENE V ( 0078 bios)
• Intel i7-3770K CPU OC'ed 4500Mhz
• Nvidia GTX480, WHQL Win7 64bit drivers
• Western Digital 1TB Green Caviar Green HDD
• Corsair AX 1200W PSU

 

 

Speeds we tested at are:

• 1600Mhs C9-9-9-27 2T Command Rate 1.5Vdimm
• 1866Mhz C9-10-9-27 2T Command Rate 1.5Vdimm
• 2133Mhz C9-10-11-27 2T Command Rate 1.65Vdimm
• 2400Mhz C9-11-11-27 2T Command Rate 1.62Vdimm
• 2666Mhz C10-12-12-31 2T Command Rate 1.65Vdimm
• 2800Mhz C11-13-13-31 2T Command rate 1.675Vdimm

Hopefully the above speeds will give you a good impression what to expect when you cash out for high speed dimms, this versus the more affordable mainstream versions. Since we are at Madshrimps we couldn't resist to push a bit further and run our entire test suite at 2800Mhz. Sadly with our review sample TRCD 12 was not gonna happen at these blistering speeds. Not even at 1.85Vdimm (not recommended for daily usage ) when we tried to run 2800Mhz C10-12-12-35 2T. But the in the charts listed 2800Mhz C11-13-13-31 is HyperPI 32M and 8 instances of 1024Mb HCI Memtest stable.

Time to turn up the heat and see the scores...

Time to find out how this stunning looking high end Corsair RAM kit performs. Take note that all the speed results of sub 2666Mhz are simulated Platinum versions. Real world performance can vary, depending on the IC's being used.

To start off we are utilising Superpi 32M, a favourite amongst benchers. Ideal to show off their tweaking skills, RAM clocks and insane timings. A nice scaling can be observed, especially when comparing 1600Mhz C9, a commonly used RAM frequency, versus the blistering 2666Mhz speeds. A whopping 16 seconds difference thanks to the massive increase in bandwith, Wprime 1024 is a mere CPU test and the results remain within the margin of error. Why include it then you might ask ? Just to show that not all applications benefit from faster RAM.

 

 

The AIDA bandwith test shows where the gain in SuperPi 32M comes from: the READ and COPY results get a massive boost from higher clock frequencies. Likewise a drop in latency is observed from 34ns at 1600Mhz to 30ns at 2800mhz. The big difference in Copy at 2400mhz versus higher clocks is not a fluke. The tertiary tWWSR timing is set a bit too loose by the ASUS Gene V board ( this setting is at 7 for 2666Mhz and higher, while being 5 at 2400mhz )

 

 

Y-cruncher, a multi threaded SuperPi version that stresses the IMC and the CPU cores hard. We've already observed Prime95 and OCCT stable setups crash during this test. Maybe a nice addition to the stability test suite of some of our readers. Light scaling is visible, but not as heavily as with the single threaded version.

 

 

The AIDA CPU tests have similar outcome as with the first chart on this page. PhotoworXX, relying on bandwith, while CPU queen is a pure CPU test and it's output is hardly influenced by increased bandwith. Again as with the AIDA Bandwith test, the 2666 and 2800mhz results are partly crippled, due to the loose tertiary timings. But more on that on the next page.

 

 

 

Both our Maxon Cinebench 64Bit suites show very mild scaling with extra RAM performance. Nothing earthshattering, but an improvement is an improvement.

 

 

The X264HD encoding test is where we can truely show you how fast RAM can boost your system's performance. A healthy boost for the Pass 1 test, this with RAM clocks up to 2400mhz. With even higher RAM clocks, the scaling becomes less impressive. The Pass 2 test scales too, but the gain is minimal.

 

 

3D action time , but this time we opted to only included synthetic tests. In our previous RAM reviews it was more than clear that Games hardly run better with extra RAM bandwith. Note that we are referring to real time game playing tests; Thus not utilising integrated game benchmarks. First up is the old, but amongst benchers beloved 3DMark2001. The 3D01 scores gain big time when going from 1600 to 1866Mhz RAM frequency. But from there on the performance is hardly influenced, gaining 400 points with almost 1000mhz higher RAM frequency.

 

 

 

If we look at the outcome of the two other Futuremark synthetic benchmarks we can see that the scores hardly  budge with high RAM frequencies. Similar to what we experienced while doing the game tests in previous RAM reviews. For gaming purposes 1600mhz RAM clocks is a good balance between performance and cost.

 

Each platform has it's own peculiarities regarding how to fine tune and get the most out of your hardware. With Ivy Bridge the tertiary timings are crucial regarding booting at high clocks and matter even more when starting to bench. The ASUS bios is as usual well dialed in, this to enhance RAM compatibility.  The ROG boards however get an extra treatment. The ASUS engineers did most of the hard work for it's end users by giving us preset RAM profiles. Just select the type of profile that suits your RAM kit and your needs, then start to fine tune from that starting point.

Now let's see what the board's bios sets when we load the Corsair XMP profile. The result is shown below in the in the first untweaked Mem TweakIt screenshot.

 

 

As we pointed out with a few tests on the previous page, some scores were slower at 2666Mhz then at 2400mhz. This is partly our fault, as we only change the main timings and let the board dial in the secondary and tertiary timings. We opted for this approach as this is probably what most end users will do too.

Very important for maximum performance are the tRRSR and tWWSR values. For daily usage a value of 5 is fine. However benchers will try to set 4. Take notice that this slight change could require far more Vdimm voltage. Secondly not all RAM kits remain stable with the 4 value selected for both the tRRSR and tWWSR timings.

Command Rate 2T also gives a little bit less performance then when utilising 1T Command Rate. Similar and regularly used is a far tighter TRFC setting. The XMP profile sets 214, but the kit is perfectly stable at 174 or even less. How we slightly tweaked the XMP profile settings in the quest for more performance can be seen in the below screenshot.

 

 

Most important changes we made are:

• Locking tRSSR and tWWSR at 5
• Command rate at 1T iso 2T
• Tightening TRFC to 147 iso 214
• Some slight secondary timing adjustments

How does it affect the RAM or oveal system performance ? We redid a few simple tests, which you can easily replicate. First up is SuperPI 32M, almost a 2 second gain and believe me there is more to dial in via the tertiary timings:)

 

 

The AIDA Bandwith test shows the amazing performance at the COPY test. All achieved by tightening the tWWSR setting from 7 to 5. Write and Read performance achieve a minimal gain.

 

 

The CPU Queen test, which is close to zero bandwith dependant, logically hardly improves. But just look at the PhotoworXX outcome.

 

 

 

The multithreaded Y-Cruncher stress test completes almost 3secs faster. The Cinebench Release 11 scores remain almost identical.

 

 

 

Even though the encoding test seemed already sort of maxed out at 2400Mhz. There's still the option to tweak for some extra performance. Maybe a bit time consuming, yet making a nice difference. Futuremark's 3DMark11 PhysX test increases with 200 points. However the overal score gains far less.

 

This is my first 16GB high end kit at the M lab and I must say I was impressed how smooth it all went on the ASUS board. On my Gigabyte Z77-UD5H, and honestly with all the best effort in this world, I couldn't even get the rig to post. Big credits to the ASUS bios team.

Now let's push the voltages a bit more.  With the Vdimm at 1.725V allowed us to bench and run pi 32M at 2800Mhz C11-13-12-30 1T Command Rate. But ofcourse who is gonna bench Sandy or Ivy Bridge with 16GB of ram installed ?

 

So we removed two dimms and went down for a trip to the maximum IMC lane of our i7-3770K CPU. Remember this is all done on air. 8GB 2900Mhz C11-13-13-30 1T at 1.725Vdimm is not bad at all, for non enthousiast targeted sticks :)

 

But how about X79 ? Intel's high end platform, where we can properly utilise quad channel kits. Finding a Sandy Bridge-E CPU who's IMC can handle the Corsair's 2666Mhz box rated speeds is tough. Sadly this is the case with our Intel i7-3960X ES sample. It's IMC runs on the edge already sub 2500mhz.  Only option then: downclocking time and try set the timings as tight as possible. We stumbled upon a brick wall around 2450Mhz C9-11-11-28 command rate 1T. No matter how loose the timings were set, no matter the volts we pumped through the Dimms and/or CPU/IMC related components, upping the Bclock for more RAM MHz ran far from stable. 2450MhzC9 at 1.65Vdimm is again a pretty smooth ride in my book.

 

 

Finally a forum member asked, how do they perform on X58 ? Since you already have to pay a premium price, we opted to use the 4 Dimms iso the regular preferred triple channel setup for socket 1366 platforms. How does 16GB running at 2100Mhz C9-10-10-27 2T at 1.685Vdimm and even HCI Memtest stable at 9-10-9 2T at 1.745Vdimm sound ? Madshrimps is pretty sure, if we had a better clocking Gulftown CPU, that the Corsair Platinum kit could have easily done better. The i7 970 ES CPU we have at the [M] lab has got uber uncore problems over 3200Mhz.

 

Sadly no AMD FX results as the Crosshair V Formula went up to hardware heaven during a bench session.

Overclocking results can vary and are mainly dependant on several factors : the RAM kit you get, your CPU's IMC, the used motherboard and more important it's bios. Finally don't forget the little man behind the keyboard ! Ofcourse there is always some luck involved. Just for a comparison: we couldn't achieve stability with TRCD set at 12, once running speeds at or above 2800Mhz. Indonesian overclocker Hazzan has run his 16Gb kit at 2800Mhz 10-12-12-31 1T at just 1.68Vdimm.

We quickly tested our kit for max CPU-Z validation and came up with a nice 3010Mhz at C11-14-14-35 2T with 2 sticks installed.

Well there you have it ladies and gentlemen. Corsair is back in the game, this time not with another Vengeance kit, but the stunning looking Dominator Platinum series. It remains a matter of personal taste, but I still have to encounter one person that dissaproves of how they look. Of course looks are great, but some people even want more bling bling or added features. Corsair has got you covered, as the optional LED Chrome bars might persuade case modders to opt for these rams. The LINKability in my book however remains debatable.

Performance wise the 16GB 2666C10 kit is truely high end stuff. Binned Samsung based IC's are being used, combining the best out of two worlds: high speeds together with pretty tight timings. A benchers' delight as most X79 and Z77 boards are optimised for these Samsung ICs. Less and less motherboard bioses are dialed in for the older PSC, Hyper and BBSE IC's. Logic as these RAM kits are getting pretty rare.

 

 

Secondly the compatibility with previous Intel motherboard chipsets is pretty decent too. Usually it will be the CPU's memory controller which will run into a brick wall first, so before you can maximise these Corsair 2666C10's. Once our AMD FM2 mobo's start to roll in we can test these precious Dimms again on AMD CPUs.

These 2666C10's are magnificent sticks: perfect for those that want high quantity, but not wanting to compromise performance at any level. The new stunning look is an added bonus and will be a real showstopper in any rig. Even though the Platinum series were not intented to be, but this version is pretty perfect for benchers too. As high speeds with decent timings are so easy to achieve. Maybe only drawback for the previously mentioned crowd is that these 2666C10's are only available in 16GB kit, an 8GB version would be far more appropriate for most.

To sum it up: The brand new Dominator Platinum 2666C10 is here not only to impress, but also to knock your socks off performance wise. Albeit for some with a steep price tag of around 500 dollars at the Corsair Web shop, but that price is more than justified looking at the specifications and performance they bring. If you want one of the best performing kits around combined with a killer look, then these should be on the top of your shopping list !

 

 

We hardly ever give two awards, but these Platinums have impressed your humble reviewer and fellow benchers of the MADSHRIMPS crew big time.

 

PROS:

• Binned Samsung ICs: high speeds yet tight timings
• Stunning heatspreader design
• Overclock/Tweakability

CONS:

• 8Gb versions please for the high end kits

 

A big thanks to Gareth Ogden from Corsair for the sending us Platinum 2666C10 kit

 

 

And the crew from Tones for the hardware support during all these years.