You are not registered yet. Please click here to register!
New D9 are weeker and weeker... is that true? - i4memory.com - different look at memory
Tags Register Blogs FAQ Members List Calendar Mark Forums Read

Notices

Reply
 
Thread Tools
Old 11-08-2007, 09:40 AM   #1
Miravo
Senior Member
Miravo's PC Specs
 
Join Date: Aug 09 2007
Location: Warsaw, Poland
Posts: 146
New D9 are weeker and weeker... is that true?

It isn't any concrete topic, but I am playing with D9 GMH / GKX long time and I notice some changes:

Vices:

1) overclocking with 3-3-3-X / 4-4-4-X is much worse then chips from 2006 year
2) Alphas, new D9 "love" very loose latency, especially TRTP (my private surveillance - on Asus P5B Del)
3) They stop scaling with voltage at ~ 2.9v

Virtues:

1) Overclocking with 5-5-5-15 is the same or even better
2) With good motherboards they don't die so quickly
3) They are chipper than old ones
4) They are more resistant for voltage above @ 2.6v

Examples (all 2.45v Asus P5B Deluxe)

Old ones:

G.skill HZ PC6400 2x1 GB @ 550 4-4-4-12 (Stable)
G.skill HZ PC8500 2x1 GB @ 575 4-4-4-12 (Stable)
G.skill ZX PC8000 2x1 GB @ 520 4-4-4-12 (Stable) - crappy ones
G.skill HZ PC6400 2x1 GB (606 week) 540 4-4-4-12 (Stable)
Team Group PC 9600 2x1 GB (one of the first sets) 585 4-4-4-12 (Stable)
Mushkin XP2 PC6400 2x1 GB @ 560 4-4-4-12 (Stable)
Goodram PRO PC6400 2x1 GB @ 530 4-4-4-12 (Stable)

New:

G.skill HK PC8500 (new model) 2x1 GB @ 540 4-4-4-12 (Stable)
G.skill HZ PC6400 (new heatsinks) 2x1 GB @ 550 4-4-4-12 (Stable)
Mushkin XP2 PC6400 (new heatsinks) 2x1 GB @ 540 4-4-4-12 (Stable)
Mushkin XP2 PC8500 (from the period of price cut) 2x1 GB @ 530 4-4-4-12 (Stable)
Team Group PC9600 2x1 GB (newer than old one) @ 570 4-4-4-12 (Stable)
Goodram PRO PC8900 2x512 MB @ 570 4-4-4-12 (Stable)
Goodram PRO PC7300 2x512 MB @ 550 4-4-4-12 (Stable)


With more volts (from bios / various mobos)


Old:

Very old G.skill HZ PC6400 2x1 GB @ 560 4-4-4-12 (SPI16M) @ 2.75v - died
G.skill HZ PC8500 2x1 GB 2.6 @ 610 4-4-4-12 (Stable!) scaling up to 3.1v almost died
Team Group PC9600 2x1 GB 2.75v @ 635 4-4-4-12 (SPI 32M) - died/scaling up to 3.425
Mushkin XP2 PC6400 2x1 GB 2.6v @ 600 4-4-4-12 (Stable) scaling up to 3.1v / problems with booting from. 2.7v
Goodram PRO PC6400 2x1 GB 2.67v @ 567 4-4-4-12 (Stable) (no force to eat more volts...)

Old D9 started dyeing from 2.65v permanently

New:
G.skill HZ PC6400 2x1 GB (new heatsinks) @ 2.9v (yeah, brainpower rox!) @ 485 3-3-3-12 (SPI32M)
Mushkin XP2 PC6400 2x1 GB @ 2.9v 600 4-4-4-12 (stable)
Goodram PRO PC 8900 2x512 MB @ 2.9v @ 650 4-4-4-12 (SPI32M)
Goodram PRO PC7300 2x512 @ 2.9v @ 600 4-4-4-12 (SPI32M)
Mushkin XP2 PC8500 2x1 GB @ 2.7v 560 4-4-4-12 (SPI32M)
G.skill HK PC8500 2x1 GB @ 2.9v @ 600 4-4-4-12 (stable)

Mushkin XP2 PC6400 & G.skill HZ PC6400 (New heatsinks) I was bulling with 2.9v for almost 3 weeks (minimum 3hours per day) and until now they are completely stable and overclock the same. Old D9 were changing after few hours of playing with hazardous voltage

Jot down examples (results) and write your opinions please

With regards
Michael
Miravo is offline   Reply With Quote
Old 11-08-2007, 01:49 PM   #2
eva2000
Administrator
eva2000's PC Specs
 
Join Date: Jul 22 2004
Location: Brisbane, Australia
Posts: 23,010
twitter.com/i4memorycom Facebook Page livestream.com/i4memorycom
Blog Entries: 42
Re: New D9 are weeker and weeker... is that true?

I think it depends on what motherboard you use as well..

On the base reference system P5B Deluxe 1101 bios I have tested all
memory including 10000C5DF and it ended up like this:

2x1GB Corsair 10000C5DF
- maxed 570mhz 4-4-4-5 with 6-42-11-11-11-11 at 2.45v
- maxed 600mhz 5-5-5-15 with 6-42-11-13-11-11 at 2.45v (very very
loose subtimings)
- no tighter than 5-5-5-15 or higher than 600mhz regardless of vdimm used.
- needed very loose subtimings to get stable

2x1GB Teamgroup 6400C3 on mid 2006 D9GMH
- maxed 580mhz 4-4-4-5 with 3-30-3-10-10-10 at 2.45v
- maxed 595mhz 4-4-4-5 with 6-42-10-10-10-10 at 2.63v
- no where close to 2x1GB OCZ PC2-8000 Titanium Alpha VX2 on Asus 680i Striker in terms of handling tight subtimings

2x1GB OCZ PC2-8000 Titanium Alpha VX2 D9GKX or D9GMH
- maxed 565mhz 4-4-4-4 with 6-35-10-10-10-10 at 2.45v
- maxed 585mhz 4-4-4-4 with 6-42-410-8-10 at 2.60v
- but on Asus 680i Striker this pair is easily best at 650mhz 5-3-3-3 tight subtimings capable

2x1GB G.Skill 6400PHU2 HZ D9GMH 629 batch
- 555mhz 4-4-4-4 with 3-30-3-10-8-10 at 2.45v
- 600mhz 5-4-4-4 with 3-30-3-10-8-10 at 2.60v

2x1GB SuperTalent T1000UB1G5 D9GKX or D9GMH
- 533mhz 4-4-4-4 with 3-30-3-9-6-9 at 2.45v
- 605mhz 5-5-4-5 with 3-30-3-10-6-10 at 2.35
- 600mhz 5-5-5-15 with 3-30-3-10-8-10 at 2.3v
- 600mhz 5-4-4-4 with 3-30-3-10-6-10 at 2.62v

2x1GB OCZ PC2-9200 FlexXLC D9GKX or D9GMH
- 600mhz 5-5-4-5 with 4-25-11-11-11-11 at 2.45v
- 583mhz 5-5-4-5 with 5-25-9-11-9-11 at 2.35v
- 540mhz 4-4-4-12 with 6-42-10-11-10-10-10 at 2.45v
- no tighter than 5-5-4-5 at higher end and 4-4-4-12 ceiling around 540mhz
- 2nd best on Asus 680i Striker doing 650-670mhz 5-5-5-15
- But fly on Asus P5K Deluxe with tight 2-25-2-2-2 sub timings capable but same 540mhz 4-4-4-x limit

2x1GB Crucial Ballistix 8500 D9GKX or D9GMH
- 620mhz 5-5-5-15 with 6-42-10-10-10-10 at 2.35v (But on Asus P5K Deluxe 651-661mhz 5-5-5-15 at <2.475v isn't a problem see http://i4memory.com/showthread.php?t=4310)
- yet to test all Ballistix/Tracer 8500 pairs yet

It does look like some of the higher 5-5-5-x capable Micron D9 these days are exhibiting same characteristics, of low 4-4-4-x clocking and restricted to 5-5-5-x or 5-5-4-x clocking with fairly loose subtimings.

The Corsair 10000C5DF, Crucial Ballistix 8500, OCZ PC2-9200 FlexXLC and some extent Super Talent T1000UB1G5 all have similar behavioural characteristics in that:

1. they are really restricted to 5-5-x-x timings at high clocks
2. very low 4-4-4-x oc'ing head room

But differentiate themselves in how they handle tight subtimings (or what P5B Deluxe considers default 6-42-10-10-10-10/11 with later bios). SuperTalent and Crucial handle very tight subtimings while 10000C5DF and OCZ 9200 FlexXLC i have need very loose subtimings.

Maybe that's the reason Corsairs 8888C4D kits are rare as it's harder these days to find Micron D9 modules to do higher 4-4-4-x clocks at decent voltages ?

Last edited by eva2000; 11-08-2007 at 02:28 PM.
eva2000 is offline   Reply With Quote
Reply

Bookmarks

Tags
true, weeker


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools



All times are GMT +11. The time now is 05:15 PM.

no new posts