Wednesday, June 7th 2017

G.SKILL Memory Breaks DDR4 5.5 GHz World Record Barrier

G.SKILL International Enterprise Co., Ltd., the world's leading manufacturer of extreme performance memory and gaming peripherals, is excited to announce the DDR4 frequency world record at DDR4-5500MHz. This astonishing achievement was accomplished by the Taiwanese professional overclocker, Toppc, using G.SKILL DDR4 built with Samsung 8Gb ICs, on the latest MSI X299 GAMING PRO CARBON AC motherboard and Intel Core X-series processor.

Just last year, one week before Computex 2016, the renowned Taiwanese overclocker, Toppc, broke the DDR4 5GHz barrier using G.SKILL memory and MSI Z170I GAMING PRO AC motherboard. A year after, he raise the bar once again and became the first ever to push DDR4 speed to 5.5GHz under liquid nitrogen cooling. This tenacity demonstrates his overclocker's spirit of pursuing faster speed and greater performance of the latest computer hardware. This record has been validated by HWBOT.
For more information, visit this page.

"DDR4 5.5GHz has been our next target after we achieved DDR4 5GHz last year. We are extremely excited to finally make it happen together with Samsung components, MSI X299 motherboard, and Intel Core X-series processor," says Tequila Huang, Corporate Vice President, G.SKILL International. "We are seeing amazing overclocking potential for these newly released hardware and we believe that more overclocking benchmark records will be achieved very soon by professional overclockers worldwide."
Add your own comment

20 Comments on G.SKILL Memory Breaks DDR4 5.5 GHz World Record Barrier

#1
Nuckles56
I wonder how much bandwidth that RAM had and also what timings were used for such a feat. Also it is impressive to have RAM running at a faster speed than 99.9% (and how ever many extra .999s are needed) of CPUs on this planet
Posted on Reply
#2
Duality92
Nuckles56 said:
I wonder how much bandwidth that RAM had and also what timings were used for such a feat. Also it is impressive to have RAM running at a faster speed than 99.9% (and how ever many extra .999s are needed) of CPUs on this planet
You can see what timings (at least primaries) where used, in the validation link.

http://valid.x86.fr/2cbukj

21-31-31-63-2
Posted on Reply
#3
-The_Mask-
It's 2750MHz, not 5500MHz. You can call it 2750MHz or 5500MT/s, but 5500MHz is way off.
Posted on Reply
#4
Duality92
-The_Mask- said:
It's 2750MHz, not 5500MHz. You can call it 2750MHz or 5500MT/s, but 5500MHz is way off.
DDR = double data rate, = 2*2750 = 5500 MHz.

Check your own memory in CPU-Z for example. If you have 1600 it will say 800, if you have 2400, it will say 1200 and so on.
Posted on Reply
#5
ironcerealbox
-The_Mask- said:
It's 2750MHz, not 5500MHz. You can call it 2750MHz or 5500MT/s, but 5500MHz is way off.
DDR man, c'mon... 2750*2 = 5500.
Posted on Reply
#6
Crazy zookeepster
Talk about having to dump the voltage and clocks to inherently supply power to the ram modules lol. This is an amazing feat in terms of memory. When DDR4 launched, 3200 MHz was awe inspiring and look now, 5500 MHz. Well done Toppc
Posted on Reply
#9
Manu_PT
Can it run at a mere 3200mhz on AMD chipset tho? xD
Posted on Reply
#10
Countryside
Manu_PT said:
Can it run at a mere 3200mhz on AMD chipset tho? xD
Cmon Please at the moment 3200 is no problem.
Posted on Reply
#11
Manu_PT
Not every kit runs at 3200mhz, trust me. People talk about it like it is assured they can do the 3200mhz, and it isn´t.
Posted on Reply
#12
JunkBear
Can I Do it with Kingston Valueram?
Posted on Reply
#13
Prima.Vera
And there I thought this was done on air....
Posted on Reply
#14
Enterprise24
Quiet unbelieveable this is done on x299 pro carbon not APEX or OCF or even XPower. I have a bad experience on z170 pro carbon which advertise mb can handled 3600mhz but I can't go beyond 3333mhz with stability.
Posted on Reply
#15
btarunr
Editor & Senior Moderator
JunkBear said:
Can I Do it with Kingston Valueram?
That's overkill. Use a potato.
Posted on Reply
#16
-The_Mask-
Duality92 said:
DDR = double data rate, = 2*2750 = 5500 MHz.

Check your own memory in CPU-Z for example. If you have 1600 it will say 800, if you have 2400, it will say 1200 and so on.
ironcerealbox said:
DDR man, c'mon... 2750*2 = 5500.
Lol, come one, you two can't be seriously... Just read my post again.

That's like saying: "I overclocked my quadcore CPU to 16GHz!!!" (because 4 core's all at 4GHz) :kookoo:


btarunr said:
That's overkill. Use a potato.
Could you fix 5500MHz in the article btarunr?
Posted on Reply
#17
Crazy zookeepster
-The_Mask- said:
Lol, come one, you two can't be seriously... Just read my post again.

That's like saying: "I overclocked my quadcore CPU to 16GHz!!!" (because 4 core's all at 4GHz) :kookoo:



Could you fix 5500MHz in the article btarunr?
I don't want to upset you, but it doesn't matter who is correct here (Although I believe that 2750MHz*2=5500MHz is correct), the thing is that it is reported like this world wide, and not only that, all reviewers who go over this part to reassure readers, do the exact same math. This is a different concept to all others, but frankly it is how it works.
Posted on Reply
#18
Crazy zookeepster
btarunr said:
That's overkill. Use a potato.
That's overkill. Use an olive
Posted on Reply
#19
-The_Mask-
Crazy zookeepster said:
I don't want to upset you, but it doesn't matter who is correct here (Although I believe that 2750MHz*2=5500MHz is correct),
Well it's 100% wrong and it does matter. You can't just random multiply numbers by two, that's just silly.
the thing is that it is reported like this world wide, and not only that, all reviewers who go over this part to reassure readers, do the exact same math. This is a different concept to all others, but frankly it is how it works.
Most of the time it's used right, look for example at CPU-z or HWBot as shown in the article. Only some people that don't know that much about hardware call it wrong in my experience. But of course that's no reason to do it also wrong, those people only gonna believe that they are right even if they can't be more wrong.

In the beginning of the dual core CPU's a lot of people also multiplied the CPU frequency by two, because it was a dual core. Do you also think we should go back to that? I hope not :D
Posted on Reply
#20
ironcerealbox
-The_Mask- said:
Lol, come one, you two can't be seriously... Just read my post again.

That's like saying: "I overclocked my quadcore CPU to 16GHz!!!" (because 4 core's all at 4GHz) :kookoo:



Could you fix 5500MHz in the article btarunr?
Technically, yes, you are correct in that it is 2750MHz actual clock. However:
1 - This is a semantic dispute
2 - The technical world generally accepts "x real clock * 2 = effective clock", which then reverts back to point 1
Posted on Reply
Add your own comment