Gigabyte GA-EP45-EXTREME Living Review

[Ace-a-Rue]

with my memory (GSkill DDR-6400)...my max stable is 515 FSB using 8.5X = 4382 Mhz....if i had some of that Cell-Shock, i would be zooming

[Alcibiades]

supershanks
why is the everest memory latency so high??
Is it a limitation of the ddr2 sticks?
You seem to be using the same ddr2 modules as eva2000?
How the hell did eva get cas 4 timings at those speeds is AMAZING
63-4 latency is WAY too high
with your cpu clocked at over 4Ghz your superpi 32m should be about 50-60secs faster
ddr2-8500 is rated at 1066mhz, so it can only make the task of doin' close to 600fsb @ 1:1 that much more difficult i imagine.
But ddr2 latency should be closer to 50 than 60

I'd really like to see what you can do with higher rated ddr2's if you have some.

[supershanks]

You seem to be using the same ddr2 modules as eva2000?

Yes but I'm not sure how they compare, the set I have are relatively new set CL1118Y.61 which were used in the tests. I have an older set which are CL11183.MW, but I think their not as giood as they were.

I have to admit I've not really pushed tight ram , focussing more on stability for now.

I've also got some weird results today , rerunning the tests in everest , with some faulty results :-
Memory Read

Memory Write

Memory Latency


I'm using everestultimate_build_1463_sqd4xmgr8vc.zip which is supposedly the right one with link emailed from Lavasys.

ddr2-8500 is rated at 1066mhz, so it can only make the task of doin' close to 600fsb @ 1:1 that much more difficult i imagine.
But ddr2 latency should be closer to 50 than 60

Yes , though i've run at 1200 quiet easily in the past with them, but as I said above i'm not sure of current quality.

I'd really like to see what you can do with higher rated ddr2's if you have some.

I would too , but have invested my £ in my Cellshock blue The Crucial has been bought out of retirement for this review , it usually runs in my wife's PC

[eva2000]

the 1M 564mhz 4-4-4-x is 1M stable and nothing more so you can't exactly say it's 100% stable heh but it needed way more vdimm than other boards i.e. on DFI X38 same sticks do 580mhz 4-4-4-5 at 2.4v

Originally Posted by supershanks

Yes but I'm not sure how they compare, the set I have are relatively new set CL1118Y.61 which were used in the tests. I have an older set which are CL11183.MW, but I think their not as giood as they were.

I've tried several 2x1GB Crucial Ballistix PC2-8500 non-tracer and tracer kits with -/+ 5-10mhz diff in max FSB for dual channel - all kits are double sided micron D9GMH as I have 11 of these kits

Current set i'm using is 2x1GB Crucial Ballistix PC2-8500 non tracer CL1118D.GZ

I've also tried the following in hopes of higher FSB 1:1 2.0B

2x1GB Corsair PC2-10000 Dominator D9xxx double sided
2x1GB Supertalent PC2-8000 D9GKX
2x1GB G.Skill PC2-8500CL5D PK Powerchip ICs
2x1GB Team PC3-6400 3-3-3-8 D9GHM
2x1GB OCZ PC2-9200Flex XLC D9GMH

but no luck getting stable past 560FSB 5-5-5-15/24 dual channel with my E8500 Q740A479T and even them clock for clock performance is slower than usual

one thing i noticed is EP45 Extreme F6A bios at least is ignoring the memory's spds for subtimings when set on AUTO for all the memory i tested above - probably the bios isn't mature enough to properly work with all the above memory or it's limit of my cpu at 560FSB ?

here's hoping to the next bios being much better

[supershanks]

Hi George thanks for input
Crucial set i'm using in these tests is single sided (CL1118Y.61 )
I've been using DDR3 for a couple of months , coming back to running the Crucial on here has been odd, I've been struggling to see whether the ram is the weakest link or optimising the bios settings is the issue. The CPU side feels very stable. On the X48T I used Skew control to enablre a higher FSB, so far on this board I've just left it at 700mV with no Skew. It's always been the mamory subsystem that has required attention.

The fact that you've access so many sets , as you say points to an immature bios Helps stop me getting frustrated with my ram

I'm also in the same boat as yourself I think have done 4109.66 MHz (587.09 * 7) not benchable memory unstable. Have played a bit in 570's but no stability yet.

What version of Everest are you running ?? As you saw I had some strange results with mine (everestultimate_build_1463_sqd4xmgr8vc) The new bench below was run twice 1st run memory read was low as in the 1st pic (memory Read) in my last post.

I played about today looking at tweaking up performance :-

I changed :-
Tras - 24 ---> 20
Channel A & B
Trd2rd(Different Rank)....................:Auto 8 ----> 9
Twr2wr(Different Rank)...................:Auto 8 ----> 7
Twr2rd(Different Rank)...................:Auto 7 ----> 8
Trd2wr(Same/Diff Rank)..................:Auto 9 ---->8

The 9,7,8,8,came from testing with the P45T which I've just started to test. On the P45T I have 500x8 & 7.5x550 occt stable at 3.33D using the above . On the EP45 it's not as stable as the original bench may need to slacken tras back a little.



Original

[Alcibiades]

Originally Posted by supershanks

I changed :-
Tras - 24 ---> 20
Channel A & B
Trd2rd(Different Rank)....................:Auto 8 ----> 9
Twr2wr(Different Rank)...................:Auto 8 ----> 7
Twr2rd(Different Rank)...................:Auto 7 ----> 8
Trd2wr(Same/Diff Rank)..................:Auto 9 ---->8

52.2 latency, now that's what i'm talking about, superpi32 much better
nice job supershanks
if you can do superpi32 run to end doesn't that mean your ddr2 is stable? This is what I assume, correct me if I'm wrong guys.
So maybe you need a little more juice for your e8500 for complete system stability?
But again shouldn't a successful completion of superpi32m mean ram volts/subvolts and timings are sweet?
I assume that superpi is used to confirm ram overclock stability
and prime95 is used to confirm cpu overclock stability
If prime fails I adjust cpu volts/subvolts including mch
If superpi fails I adjust ram volts/subvolts and timings
See my logic is that if you have a problem with ram settings, ie. tras too tight, then superpi32m should fail

[supershanks]

52.2 latency, now that's what i'm talking about, superpi32 much better

Ta

Well yes it's benchable stable , but not prime or OCCT as the initial shot was. So as nothing has changed with the CPU side , it's the ram that is unstable, a direct result of the above changes. It may be that playing with the Skew or sub volts or vdimm may stabalise it I've not tried yet.

[supershanks]

4025 7x575@2.00B = 1150 5,5,5,22
Got my P45T in my case and on water, so this is on air , hence the reason I've run it at 7 rather than 7.5. Quiet pleased to have got this , had thought I'd about maxed it out on stable clocks at 565

Benches


Stability


Bios

GIGABYTE GA-EP45_EXTREME
[u]MB Intelligent Tweaker(M.I.T.)
Robust Graphics Booster ...............: Turbo
CPU Clock Ratio ..........................: 7x
Fine CPU Clock Ratio.....................:.+0.0
CPU Frequency ...........................:4,020 Mhz
Clock Chip Control 
Standard Clock Control
CPU Host Clock Control..................: [Enabled]
CPU Host Frequency (Mhz) ............:575Mhz
PCI Express Frequency (Mhz) .........:103Mhz
C.I.A. 2.......................................: [Disabled]
Advanced Clock Control [Press Enter]
CPU Clock Drive...........................: 700mv
PCI Express Clock Drive.................: 700mv
CPU Clock Skew (ps)....................: 0ps
MCH Clock Skew (ps)...................: 0ps
DRAM Performance Control
Performance Enhance...................: [Turbo]
Extreme Memory Profile (X.M.P.).....: [Disabled]
(G)MCH Frequency Latch...............: [Auto]
System Memory Multiplier ..............: 2.00B
Memory Frequency (Mhz) ..............: 1,150Mhz
DRAM Timing Selectable ................: [Manual] 
Standard Timing Control
CAS Latency Time........................: 5
tRCD .........................................: 5
tRP'...........................................:  5
tRAS..........................................:  22
Advanced Timing Control [Press Enter]
tRRD...........................................: 4
tWTR..........................................: 4
tWR............................................: 8
tRFC...........................................:54
tRTP...........................................: 4
Command Rate (CMD) ....................: [Auto]
Channel A
Static tRead Value.........................:Auto 10
 tRD Phase0 Adjustmen...................:Auto 0
tRD Phase1 Adjustment...................:Auto 0
 tRD Phase2 Adjustment .................:Auto 0
 tRD Phase3 Adjustment..................:Auto 0
Trd2rd(Different Rank)....................:Auto 8
Twr2wr(Different Rank)...................:Auto 8
Twr2rd(Different Rank)...................:Auto 7
Trd2wr(Same/Diff Rank)..................:Auto 9
Dimm1 Clock Skew Control...............:Auto  ps
Dimm2 Clock Skew Control...............:Auto  ps

Channel B
Static tRead Value.........................:Auto 10
 tRD Phase0 Adjustmen...................:Auto 0
tRD Phase1 Adjustment...................:Auto 0
 tRD Phase2 Adjustment .................:Auto 0
 tRD Phase3 Adjustment..................:Auto 0
Trd2rd(Different Rank)....................:Auto 8
Twr2wr(Different Rank)...................:Auto 8
Twr2rd(Different Rank)...................:Auto 7
Trd2wr(Same/Diff Rank)..................:Auto 9
Dimm1 Clock Skew Control...............:Auto  ps
Dimm2 Clock Skew Control...............:Auto  ps

{U]Motherboard Voltage Control{/U]
Voltage Type.………... Normal 
CPU
CPU Vcore….……….........................:1.4v
CPU Termination…..... 1.200V*.......: 1.34v
CPU PLL…………………....1.500V*.......: 1.60v
CPU Referen.…………....0.755V*.......: 0.813v..60.67% to vtt
CPU Reference2………,,0.800V*.......:0.843v...62.90% 
MCH/ICH
MCH Core…………….....1.100V...........: 1.420v
MCH Reference….…….0.800V...........; 0.768v  57.3% to vtt
MCH/DRAM Ref.…......0.900V...........: 0.990v
ICH I/O……………….....1.550V............: 1.54v
ICH Core…………...……1.100V............: 1.10v
DRAM
DRAM Voltage ……....1.800V............: 2.220v
DRAM Termination .…0.900V............: 0.99v
Channel A Reference 0.900V............: 0.99v
Channel B Reference 0.900V............: 0.99v
[b]{U]Advanced Settings[b]{U]
Limit CPUID Max. to 3.....................: [Disabled]
No-Execute Memory Protect............: [Disabled]
CPU Enhanced Halt (C1E)................: [Disabled]
C2/C2E State Support....................: [Disabled]
x C4/C4E State Support..................: [Disabled]
CPU Thermal Monitor 2(TM2) ...........: [Enabled]
CPU EIST Function.........................: [Disabled]
Virtualization Technology................: [Disabled]
Note: * refers to 45nm CPU

[eva2000]

Still figuring out how to squeeze some more clock for clock performance on EP45 Extreme, getting somewhere but still no where as fast as P35/X38/X48 right now.





Full report Gigabyte GA-EP45 Extreme rev1.0 Review - View Single Post

[eva2000]

more results - starting to get somewhere

7.5x533FSB 2.40B 640Mhz 5-5-5-15

tRD 9


tRD 9 + Tweak 1


tRD 9 + Tweak 1 + 2


Super Pi 1M with Tweak 1 + 2


Super Pi 32M with Tweak 1 + 2


32M time still around 20-25 seconds slower than P35/X38/X48 but getting there

rest of it at Gigabyte GA-EP45 Extreme rev1.0 Review - View Single Post