DDR3 Speed Check
The original purpose of this exercise was to check out how the new RipjawsZ series compared to the older RipjawsX, but as I had access to some other modules I thought I would chuck those into the mix as well
So on the bench I have
2x2GB Hynix 1333 ECC CL9-9-9-24-1T
2x2GB Kingston Lovo 1600 CL9-9-9-24-1T
2x4GB G.Skill RipjawsX 2133 CL11-11-11-30-2T
4x4GB G.Skill RipjawsZ 2133 CL9-11-10-28-2T
I will apologise for the photographs now, my trusty old Kodak DC290 has finally died on me after more than a decade of service
I turned it on, the lens came out and that was it, it just shut down and refuses to turn on again
So I have had to use my phone camera for this one, sorry
Test setup
Motherboard: MSI Big Bang Marshall
Processor: Intel Core i7-2600K @ Stock and OC @ 4.6GHz
Graphics card: MSI GTX580 Lightning
Cooling: Swiftech Apogee XT rev. 2 CPU water block
Power supply: Silverstone SST-ST1000P 1000watt modular PSU
Operating system: Windows 7 64Bit Ultimate
HDD: 2 x Western Digital Black WD1602ABKS 160GB in RAID0
Overclocking
I decided to keep to a standard overclock here, upping the bus speed to 103.3 and the CPU multiplier to 45, I could have taken the bus speed higher but I didn’t want to clock the memory to the point where I would have to alter the timings or voltage
This takes the CPU to 4648MHz and the 2133 modules to 2200MHz
The 1333 modules rise to 1376MHz and the 1600 modules to 1632MHz
PC Mark Vantage x64
The RipjawsZ suffers a little on the video tests here but make up for it in the CPU tests and while the RipjawsX scores better in the video tests they fall behind in the CPU tests
Overclock results
The overclock results show the same trend, with the slower modules keeping a fare pace going
A poor result for the RipjawsZ kit highlighted by the improved performance of the RipjawsX when overclocked
Aida 64 Extreme Memory Benchmark
The RipjawsZ take a clear lead here with the RipjawsX performing poorly compared to the lower speed modules and the Lovo putting in another good result
Overclock
The improvement found in the RipjawsX on the overclock is quite impressive, taking it clear of the slower kits and near on a par with the RipjawsZ
Cinebench 11.5
Dependant on the CPU and GPU the overclocked CPU is the booster here
Although the CPU tests show very little, the RipjawsX again trail the Lovo @ 1600
Overclock
The RipjawsZ are still in front but again the overclock of the RipjawsX shows a big performance increase, the Lovo suffers most on this one
Sisoft Sandra
Memory Bandwidth
The RipjawsZ simply stole this test, running away from the others
The poor little brother RipjawsX got beat on by the slower kits here
Overclock
Yet again the overclock gives a massive performance boost to the RipjawsX
While the other kits received an almost linear improvement, the RipjawsX kit stole a march on them
Sisoft Sandra
Cache and memory bandwidth
The Cache test loads the CPU cache as well as the memory to show how the memory performance affects the CPU
The RipjawsX kit seems to suffer badly with its CL11 timings at stock speeds
Even the 1333 kit is ahead at this point, again
Overclock
The RipjawsZ are still the leaders here but the massive improvement in the RipjawsX no longer surprises me, going from a poor 4th to strong 2nd
wPrime 2.07
wPrime focuses more on CPU performance
The 32m test doesn’t show much, but any improvement on the overclock score here is a bonus
The RipjawsX puts in a strong performance in this test with the RipjawsZ suffering the most
Overclock
The same story with the overclock
The RipjawsX fastest in the long test but the RipjawsZ fall behind again
MaxxMEM
MaxxMEM shows similar results to Aida64 and Sandra in that the RipjawsX fall by the wayside here, the RipjawsZ again stealing the board and the Lovo in front of the RipjawsX
Overclock
The RipjawsX make the most of the overclock yet again by hanging on to the shirt tails of the RipjawsZ and passing the slower modules
Latencies (Ns)
Aida64 Extreme, Sisoft Sandra and MaxxMEM each include a latency test in their repertoire
Unfortunately, with their smaller numbers, including latency scores with the other big scores often leaves them unseen
Grouped together you get a much better idea of what’s going on
The RipjawsZ leading with the Lovo next in line has been a common sight throughout the test and nothing changes here
The RipjawsX kit really suffering for the CL11 timings all along, the pattern repeats itself
Overclock
The results show a more linear pattern here reflecting the speed of the modules, the RipjawsX performance increase in the overclock throughout the tests has been an enjoyable surprise
Conclusion
Going back to the original purpose of this exercise, which was to compare the RipjawsZ with the RipjawsX, I would have to say that the RipjawsZ are the overall performance winners
As a DDR3 speed check, with the exception of the performance by the RipjawsX at stock speed, the performance scale is pretty linear
The RipjawsZ fared better than I imagined they would, with the exception of PCMark Vantage they pretty much ruled the roost
The biggest surprise was the performance increase gained by the RipjawsX kit when overclocked to 2200MHz
Every test showed a massive boost and at times they equalled the RipjawsZ, they were certainly in the same ball park
The Kingston Lovo 1600 put in a strong performance beating the RipjawsX on all but one of the stock tests and scaling well with the overclock
The Hynix 1333 ECC also put in a good performance at stock and while it also scaled well with the overclock the speed just wasn’t there
So now I am happy with the RipjawsZ kit and I can relax and get back to Skyrim
C ya
The original purpose of this exercise was to check out how the new RipjawsZ series compared to the older RipjawsX, but as I had access to some other modules I thought I would chuck those into the mix as well
So on the bench I have
2x2GB Hynix 1333 ECC CL9-9-9-24-1T
2x2GB Kingston Lovo 1600 CL9-9-9-24-1T
2x4GB G.Skill RipjawsX 2133 CL11-11-11-30-2T
4x4GB G.Skill RipjawsZ 2133 CL9-11-10-28-2T
I will apologise for the photographs now, my trusty old Kodak DC290 has finally died on me after more than a decade of service
I turned it on, the lens came out and that was it, it just shut down and refuses to turn on again
So I have had to use my phone camera for this one, sorry
Test setup
Motherboard: MSI Big Bang Marshall
Processor: Intel Core i7-2600K @ Stock and OC @ 4.6GHz
Graphics card: MSI GTX580 Lightning
Cooling: Swiftech Apogee XT rev. 2 CPU water block
Power supply: Silverstone SST-ST1000P 1000watt modular PSU
Operating system: Windows 7 64Bit Ultimate
HDD: 2 x Western Digital Black WD1602ABKS 160GB in RAID0
Overclocking
I decided to keep to a standard overclock here, upping the bus speed to 103.3 and the CPU multiplier to 45, I could have taken the bus speed higher but I didn’t want to clock the memory to the point where I would have to alter the timings or voltage
This takes the CPU to 4648MHz and the 2133 modules to 2200MHz
The 1333 modules rise to 1376MHz and the 1600 modules to 1632MHz
PC Mark Vantage x64
The RipjawsZ suffers a little on the video tests here but make up for it in the CPU tests and while the RipjawsX scores better in the video tests they fall behind in the CPU tests
Overclock results
The overclock results show the same trend, with the slower modules keeping a fare pace going
A poor result for the RipjawsZ kit highlighted by the improved performance of the RipjawsX when overclocked
Aida 64 Extreme Memory Benchmark
The RipjawsZ take a clear lead here with the RipjawsX performing poorly compared to the lower speed modules and the Lovo putting in another good result
Overclock
The improvement found in the RipjawsX on the overclock is quite impressive, taking it clear of the slower kits and near on a par with the RipjawsZ
Cinebench 11.5
Dependant on the CPU and GPU the overclocked CPU is the booster here
Although the CPU tests show very little, the RipjawsX again trail the Lovo @ 1600
Overclock
The RipjawsZ are still in front but again the overclock of the RipjawsX shows a big performance increase, the Lovo suffers most on this one
Sisoft Sandra
Memory Bandwidth
The RipjawsZ simply stole this test, running away from the others
The poor little brother RipjawsX got beat on by the slower kits here
Overclock
Yet again the overclock gives a massive performance boost to the RipjawsX
While the other kits received an almost linear improvement, the RipjawsX kit stole a march on them
Sisoft Sandra
Cache and memory bandwidth
The Cache test loads the CPU cache as well as the memory to show how the memory performance affects the CPU
The RipjawsX kit seems to suffer badly with its CL11 timings at stock speeds
Even the 1333 kit is ahead at this point, again
Overclock
The RipjawsZ are still the leaders here but the massive improvement in the RipjawsX no longer surprises me, going from a poor 4th to strong 2nd
wPrime 2.07
wPrime focuses more on CPU performance
The 32m test doesn’t show much, but any improvement on the overclock score here is a bonus
The RipjawsX puts in a strong performance in this test with the RipjawsZ suffering the most
Overclock
The same story with the overclock
The RipjawsX fastest in the long test but the RipjawsZ fall behind again
MaxxMEM
MaxxMEM shows similar results to Aida64 and Sandra in that the RipjawsX fall by the wayside here, the RipjawsZ again stealing the board and the Lovo in front of the RipjawsX
Overclock
The RipjawsX make the most of the overclock yet again by hanging on to the shirt tails of the RipjawsZ and passing the slower modules
Latencies (Ns)
Aida64 Extreme, Sisoft Sandra and MaxxMEM each include a latency test in their repertoire
Unfortunately, with their smaller numbers, including latency scores with the other big scores often leaves them unseen
Grouped together you get a much better idea of what’s going on
The RipjawsZ leading with the Lovo next in line has been a common sight throughout the test and nothing changes here
The RipjawsX kit really suffering for the CL11 timings all along, the pattern repeats itself
Overclock
The results show a more linear pattern here reflecting the speed of the modules, the RipjawsX performance increase in the overclock throughout the tests has been an enjoyable surprise
Conclusion
Going back to the original purpose of this exercise, which was to compare the RipjawsZ with the RipjawsX, I would have to say that the RipjawsZ are the overall performance winners
As a DDR3 speed check, with the exception of the performance by the RipjawsX at stock speed, the performance scale is pretty linear
The RipjawsZ fared better than I imagined they would, with the exception of PCMark Vantage they pretty much ruled the roost
The biggest surprise was the performance increase gained by the RipjawsX kit when overclocked to 2200MHz
Every test showed a massive boost and at times they equalled the RipjawsZ, they were certainly in the same ball park
The Kingston Lovo 1600 put in a strong performance beating the RipjawsX on all but one of the stock tests and scaling well with the overclock
The Hynix 1333 ECC also put in a good performance at stock and while it also scaled well with the overclock the speed just wasn’t there
So now I am happy with the RipjawsZ kit and I can relax and get back to Skyrim
C ya
