PQI Turbo memory is here to stay, but is it worth your hard earned money? We take a look at the best performing PC-3200 Memory PQI has to offer and found it stable with over 150MHz of head room!
Once I removed the heat spreaders I found one of the newest Samsung IC's on the market for high speed memory modules -- Samsung TCCD Revision F's. The full part number was found to be K4H560838F-TCCD. For those of you who don't know, these IC's from Samsung are in the TSOP form factor and are rated by samsung at 4ns.
Great review............ That's some really good ram, I would have liked to see how well it performed in a A64 system though as alot of the socket 939 VIA boards will hit a 250-260 FSB with the A643500.
yeah, if you are lucky enough to get one that works right!!! Most of the socket 939 boards have AGP/PCI locks that aren't functional. Actually, from what I've read, and from personal experience, all the 939 boards have at least one major flaw.
Yeah, but all I got right now is my 3800+ and it can't hit 250-260FSB I haven't checked on my other 939 CPU's yet, so I might give them a test drive real quick.
my ASUS was a v1.2, and almost none of those had working locks. The rumor was that the ASUS boards required a hardware revision, while the ABIT was a software revision. I just got back my ASUS v2.0, we'll see how it does, but it does have working locks....only now I need to go buy a wireless adapter because ASUS decided not to include it in the v2.0
As to the MSI boards, they are having the same problem as the ASUS, both their VIA and their NVidia boards supposedly have lock issues.
Last edited by LVCapo on Wed Sep 29, 2004 8:06 pm, edited 1 time in total.
I have the ASUS board with a working PCI bus lock, but PJ (the LR reviewer) has had the board for about 2 months and I don't have it here in the lab. I need to get the board back from him, but he isn't done reviewing it yet.
This is the main reason I never buy the first release of any board...........they never work quite right. just like my NF7-S, I waited till ver 2 came out before I bought it.
The problem I'm having now with the Rev2 board is that it won't run the HTT at 1000x2, only 800+x2, so I'm sitting at 212FSB and instead of running about 1050 on the HTT I'm only getting about 870. I think it has to do with the BIOS version of 1005, once I get online I'll try to upgrade the BIOS and see if that makes a difference, but I'm afraid its the lesser of two evils, the later BIOS doesn't allow for the PCI lock, only version 1005.
Just a little heads up, I've now been through 2 different sets of this crap ram and it still will not work correctly in my A64 system. The first set I had would not even run at the rated timings on any system I had, at 2-2-2-5 it got over 45,000 errors in 1 loop of memtest and the best timing I could get error free was 2-3-3-6. Sort of makes me think that the stuff reviewers is hand picked. I then got a replacement set of the memory (took over 2 weeks from PQI) and it wasn't even close enough to run in my A64 system in dual channel mode, it would run fine in single channel mode or 1 stick at a time but all I got was erros and non-post in dual mode. On a positive note, it did run OK in my old socket A NF7-S, but you won't see me recomending this ram to anybody I know.
eh i've got it and it works great Ran fine in a 754 system at stock 2-2-2-5 @2.5v. and in an intel system (865 and 915) ran great at 250 2.5-3-3-6 2.65V. Havent had a chance to try it in a 939 yet though.
It could be your memory controler. Try running a few loops of memtest on some known good ram and see what comes up. Also remember that with the integrated mem controller in the CPU it is affected by overclocking. Its possible that your pushing the mem controler beyond what it can do. Possible, but very unlikly as you know your sh*t.
"Don't open that! It's an alien planet! Is there air? You don't know!"
Xerxes wrote:eh i've got it and it works great Ran fine in a 754 system at stock 2-2-2-5 @2.5v. and in an intel system (865 and 915) ran great at 250 2.5-3-3-6 2.65V. Havent had a chance to try it in a 939 yet though.
Your 754 is only a single channel system, this ram works fine in single channel on my system also, I got it up to a 260 fsb..........but it's won't do crap in dual channel.
infinitevalence wrote:It could be your memory controler. Try running a few loops of memtest on some known good ram and see what comes up. Also remember that with the integrated mem controller in the CPU it is affected by overclocking. Its possible that your pushing the mem controler beyond what it can do. Possible, but very unlikly as you know your sh*t.
Memory controller is fine, that's what I was running before I switched out for this crap PQI junk. I did go back and retest it just to be sure it didn't take a dump on me. I ran 3 different sets of ram through the system and tested it all with memtest also. Everything from old Samsung PC2100 to Kingmax SuperRam PC2700 to Mushkin Black PC3200 all worked great with zero errors. I think the problem is that the 939 CPU's require ram that is really matched (SPD and so on) for it to run in dual mode and this PQI stuff is not close to being matched. One stick has SPD timings of 2.5-4-4-8 and the other has timings of 2-2-2-5. And I think that the controler checks that they are the same before the system posts, because I also tried booting with one stick, setting the timings and then installing the second stick, still didn't work. I tried it that way with both sticks, so I'm pretty much out of ideas short of sending the crap back to PQI AGAIN and and try a 3rd set.
That is odd, my sticks have the same spd timings. Winchesters are finally starting to become available locally so i might just have to try one in a neo2.
Are these babies going to cope well with my DFi NF4 sli-dr Expert? I can up the Vdimm up to 4V, how far are they going to reach, considering they reached 282Mhz at 2,85V?