Introduction:
The F2-8500CL5D-2GBPK kit is a relatively new product from G.Skill, designed to fit the needs of the gamer or enthusiast who is also on a budget. The sticks offer standard timings for a 1066MHz DDR2 kit, but at what many would consider to be a lower than average required voltage for such a kit at just 2.0 volts. In addition, the sticks boast a sturdy 8-layer PCB as well as a new, as of now undisclosed IC attached to them. As with much of the new offerings from many of the more popular memory manufacturers today, G.Skill takes a new IC and puts it to good use in a market segment that's absolutely perfect for it. Currently, despite an MSRP of $99, you can find this kit for just under $70 at Newegg, and you can only assume just as with nearly every DDR2 kit on the market that prices could even drop lower as the weeks and months go on and DDR3 becomes more popular.
The Memory:
-------------------------------------------------------------
Specifications:
Capacity: 2GB (2 x 1GB)
Speed: 1066MHz DDR2 (PC2-8500)
Latency: CL 5-5-5-15 (EPP Ready)
Voltage: 2.0 ~ 2.1 Volts
PCB: 8 Layer PCB
Warranty: Lifetime
Price: $69.99 @ Newegg
-------------------------------------------------------------
Test Setup:
Intel Core 2 Duo E6300
Abit IP35 Pro - Beta 16.B04
G.Skill F2-8500CL5D-2GBPK
BFG GeForce 7950GT 512MB
OCZ PowerStream 600w
CoolTechnica AQX MP-05 SP LE, Swiftech MCW60, Black Ice Pro III, Swiftech MCP655, 7/16" ID
Dominator triple fan cooler over memory.
-------------------------------------------------------------
The Results:
Each piece is separated by timings and voltages. I started with 4-4-3-10 and ended with 5-5-5-15 to achieve maximum stable speed.
The voltages listed are what was selected in BIOS.
-------------------------------------------------------------
Timings: 4-4-3-10-30-8-6-4-4-2T
2.0v - 419MHz (DDR 838MHz)
Dual SuperPi 32M | Everest Bandwidth + Latency
2.1v - 422MHz (DDR 844MHz)
Dual SuperPi 32M | Everest Bandwidth + Latency
2.2v - 425MHz (DDR 850MHz)
Dual SuperPi 32M | Everest Bandwidth + Latency
2.2v+ - 4XXMHz (DDR 8XXMHz)
As we can clearly see here, the sticks do not scale well at the relatively low latency of 4-4-3-10. Attempting to boot at CAS 3 with any stability was impossible above 800MHz, and tRCD of 3 was a complete no-go. These sticks belong running at CAS 5, while shooting for higher MHz with lower voltages.
Timings: 4-4-4-12-35-8-6-4-4-2T
2.0v - 458MHz (DDR 916MHz)
Dual SuperPi 32M | Everest Bandwidth + Latency
2.1v - 470MHz (DDR 940MHz)
Dual SuperPi 32M | Everest Bandwidth + Latency
2.2v - 477MHz (DDR 954MHz)
Dual SuperPi 32M | Everest Bandwidth + Latency
2.3v - 486MHz (DDR 972MHz)
Dual SuperPi 32M | Everest Bandwidth + Latency
2.4v - 492MHz (DDR 984MHz)
Dual SuperPi 32M | Everest Bandwidth + Latency
Timings: 5-5-4-15-42-10-8-4-4-2T
2.0v - 537MHz (DDR 1074MHz)
Dual SuperPi 32M | Everest Bandwidth + Latency
2.1v - 546MHz (DDR 1092MHz)
Dual SuperPi 32M | Everest Bandwidth + Latency
2.2v - 558MHz (DDR 1116MHz)
Dual SuperPi 32M | Everest Bandwidth + Latency
2.3v - 567MHz (DDR 1134MHz)
Dual SuperPi 32M | Everest Bandwidth + Latency
2.4v - 5XXMHz (DDR 1XXXMHz)
Could not gain a single MHz with 2.4v. It seems the chips had hit their limit with a tRP of 4.
Timings: 5-5-5-15-42-10-8-4-4-2T
2.0v - 591MHz (DDR 1182MHz)
Dual SuperPi 32M | Everest Bandwidth + Latency
2.1v+ - 5XXMHz (DDR 1XXXMHz)
Unfortunately, despite trying every combination of sub-timings I could, and any amount of voltage, I could not achieve any speed higher than 592MHz with any stability. The chipset began flaking out between 592-600 and would blue screen randomly during testing and lock up in memtest86+. We are simply looking at a chipset wall, not an issue with the memory. We can assume the sticks would have continued to scale reasonably well past this point if the trend continued.
-------------------------------------------------------------
Conclusion:
Well, when all is said and done, I have to say that this was not the most exciting kit to play with. I am a big fan of tighter timings and lower overall latencies with my memory, and I'd much rather run 3-3-3 @ 900MHz than 5-5-5 @ 1200MHz, but that could just be me missing my BH5.
With that said, I did have plenty of fun with this kit testing CAS 5, and I could certainly see the power within the chips to be able to reach very high frequencies with lower voltages than what we are used to with Micron chips. I think these sticks could sit very comfortably in someone's system running 5-5-5-15 at 1200MHz with only a slight bump in voltage over spec for total 24/7 stability, and very low heat output and power consumption.
On the topic of heat, it's safe to mention that these sticks will run quite cool in a system under 24/7 operation. Both while benching and after leaving my system on for several days straight, I could place my finger on the heat spreaders and hardly feel any warmth being generated at all. This should certainly interest not only people with workstations or those who crunch, but gamers who don't want to have to worry about being in the middle of an online game and having their system crash because of heat or voltage related memory failure.
These sticks face stiff competition from the ever-falling prices of Micron and ProMOS kits. And, while I feel these certainly have a price/performance edge over said kits, they may just simply be the only game in town now that Micron D9GCT/GMH/GKX seems to be disappearing from many of the products that once carried them. So perhaps this is the way the DDR2 market is heading, now that the performance focus has been place primarily on DDR3.
One thing is for sure; for those of you looking for a more reliable substitute to your auto-RMAing Micron D9's, these would most certainly be worth a look.
Thanks to G.Skill for the review sample. Thanks to you for reading.
Reply With Quote
Bookmarks