Jump to content
HWBOT Community Forums

sin0822

Members
  • Posts

    757
  • Joined

  • Last visited

Everything posted by sin0822

  1. Today we venture into some memory, the ADATA 2800C12 MFR kit. This is my first experience with MFR, and I have read all the editorials on MFR, and this is my take. I have always been the type of overclocker who enjoys getting high validations, but I also understand the need for efficiency especially when it comes to benchmarks. I will take a look at this kit, and push it on air and see how far we can get it in single DIMM, 2 DIMM, and 4 DIMM configurations. I also will do efficiency comparisons with 2DIMMs V 4DIMMs for those interested, which should take care of some of the lack of efficiency that comes from single sided kits. First off I will take a look at the kit: The gold heatspreaders are actually a darker shade of gold which is easier to match with a yellow/gold motherboard, however the contrast of gold and green is also pretty neat. The PCB is a nice black color and the PCB has 2oz of copper which means the memory module can cool its chips better than 1oz copper PCBs. The memory is single sided and held on by tape, which is better for removing the heat-spreaders if one wants to use LN2. Here is an upclose of the sticker. The timings are pretty loose and in this review I will try to use the same cas latency when possible. I used the Z87X-OC with BIOS B03, if you use this board please only use BIOS X08 or B03 for Hynix overclocking. X08 should yield better results. First I just loaded XMP and ran it with 2 DIMMs: Then I tossed in two other sticks and ran superpi again with 4 DIMMs to see if I could shave off a few milliseconds. I won about 2.2seconds out of 4 minutes, so not much, but as you can see there is a small benefit to running 4 DIMMs rather than 2DIMMs stemming from the fact that MFR is single sided. Now for 3000mhz at T1 2DIMMs: 4DIMMs. About 1 second gained, so not so much, but still a slight difference. The highest I could run superpi 1M with 2 DIMMs was at 3400mhz: Cas of 13, VDIMM of 1.8v. With 4 DIMMs and SPI 1M I was able to do 3200mhz using the same 1.8v. Max Validations: For max validation of 4 DIMMs I still only used 1.8v, I was able to do 3335mhz@4DIMMs which is pretty cool for air. Now moving to 3462mhz@2DIMMs same 1.8v: Here we have a 3581mhz@1DIMM which is not bad at all for 1.8v on air for MFR These sticks hold their own, even rated at 2800C12, they can do 3000C12/T1. If you are in the market, and an overclocker who likes to see how high they can get memory speed on the Z87/X79 platforms, then this MFR kit isn’t too shabby. I will say though, for what MFR lacks in tight timings and efficiency, it does make up for in easy fun. It wasn't hard to get this memory to these speeds, I just set vDIMM, cas to 13, command rate to T1 or T2 and let the board pick all the other timings. Certain dividers go higher than others 26.66x might not be as effective as 26.00x or 28.00x. With more BCLK some dividers might have better stability. If you are in the market for MFR and have the money to spend then this is a good kit to look into.
  2. oh damn i didn't notice that, it worked last time i used it. I think you can still register the SN on your local RMA page and they wil tell you the warranty status.
  3. yes it works, and i think for most regions, but just use the India one that is in the post.
  4. haha that one made me laugh!
  5. sin0822

    Ex Tee You!

    well its and Intel hwbot partnership dude, what you expect?
  6. here: https://www.facebook.com/teamauoc http://www.mediafire.com/?byu5xiuom51mwa4 A couple notes i want to add: To change RTLs different per channel(the board might auto do this correctly, but if it doesn't, then you will need to use fully manual mode, meaning youll have to set all timings the same per both channels(ouble work) So IMO save your profiles to USb drive as well as the profiles just so you don't have to redo it all the time. You will see two RTL and two IOL change things under manual mode, but those are just for each side of the single DIMM for that channel, so you need to set most everything same except the RTLs/IOLs per channel. (THis will be remedied to make it much easier to set both channel's RTLs/IOLs without having to set all timings twice) DRAM OVP/OCP has been worked on, and also so has vDIMM capability even on air you can do over 2.0+v, SamOCX you got your 2.4v BIOS here. This BIOS should be better with subzero(it has fully unlocked voltages). tWR still isn't working perfect, however tWR is just part of another timings tWRPDEN, so to change tWR just set something in the BIOS like 10 for tWR, and then set your tWRPDEN to 11 or 12(i forget which I used) and you will get tWR of 10. (This will be fixed soon, but it isn't impossible to change tWR effectively, its actually quite easy) tWRPDEN=tWR +/- some things. This BIOS enjoys training the memory on every boot, so from a cold boot or a hard reset expect a retrain of the memory, sometimes its best not to retrain so you might need to get a feel for it. (This will also be remedied in a future release) Dinos and Cookie i am sure will post up how they got their top memory OCs very soon. This is the First revision of the X series BIOS, second will only get better.
  7. soon i guess... where is the teamau facebook page anyways? i couldn't find it, but I am not so facebook savvy.
  8. idk money is everything with how companies and industries view things like sports.
  9. yea he will post it when he gets up it ill be like b15. Still better than B11, and there will be more updates as well. Youll get your vDIMM VRM/PWM restrictions all gone too, lilke 2.4v no OVP. stuff like that. But also should have some better compatibility with inputting settings, but you might need to use advanced manual mode more than you like. lol
  10. yea man with those clocks he should have the DDR3 frequency WR if he can pull that out of that kit.
  11. wait did they say 1 million dollars? US or NTD?
  12. yea i just saw this shit, what an abomination to reviewing, you have to put up screenshots at least. Also cas12 isn't awesome and running memory higher on retail CPUs kills CPU/cache OC. espcially at 29.33. On ES not so much tho
  13. I had a friend do it. I don't like software.
  14. See now you are smashing it like a BOSS. You make samsung look like sh*t. Maybe i should add it to the list, but i thought we made sure to put in to max out DRAM OCP/OVP? but phase control I always max out b/c that VRM doesn't need to focus on its temperature, it needs to focus on its output. Anyways doesn't matter with the next BIOS anyways. Glad you got it work goofy.
  15. What is odd is that i even had some QEH6 that loved 3v vin when vcore was only like 1.8v, it gave me 5C extra on the CB, CPU never died or degraded. How scary to try higher vin and get death while you think it would help. QEH6 should be identical to retail other than microcode. Maybe it is some BIOSes on some boards? My buddy just killed his best CPU on the M6E with 2.8v VIN and BIOS 0711. I used 0021 and 0605 with the M6E. I tested the M6E pushed 3v input voltage, and highest QE6S could do like 1.65v lol, then i moved to max vcore mode and gave it 2.5v(so like 2v) and still no degradation or death, of course i didn't use the Ln2 profiles b/c they are a bit crazy for my tastes and i just wanted to see what the hype is about.... But I also didn't enable max cache voltage mode, I always made sure that was disabled and just running normal cache volts, maybe it has something to do with the death, maybe some voltages kill. No death for me, but then again this CPU is pretty resilient.
  16. some dudes hacked my site, but I know who they are now.
  17. no one wants to list their boards and settings?
  18. most likely b/c those kinds of memory lose money for the company, and they are marketing stunts more than anything. They bid on these high end batches of ICs and then they are stuck binning them which is a big hassle and puts a lot of pressure on them. But seriously kudos for making a kit like that and selling it in public. with 4 sticks i bet the peformance would have been okay.
  19. i think maybe he was impressed by its OC capability and figured if you bought the kit you'd probably buy it for high frequency epeen like they are competing on facebook right now with, its pointless, yes.
×
×
  • Create New...