RAM timings show 5-5-5-18, not 5-5-5-15

[Cheryl_27]

Put in 2 new dual channel kits of 4 GB of Kingston HyperX KHX6400D2B1K2/4G (maxed at 8GB). Kingston Specs

Everything looks right in PC Wizard / CPU-Z, except for tRAS timing. Per Kingston specs, at 400 MHz (which it's at), timings at 1.8 v should be 5-5-5-15.

The SPDs are programmed to JEDEC standard latency 800MHz timing of 5-5-5-15 at 1.8V. Each 240-pin DIMM uses gold contact fingers and

requires +1.8V.

1.8v is the spec on this mobo & was the voltage of the OEM RAM (however, their specs were 6-6-6-18). This is an HP OEM Asus board w/ no clock settings in BIOS. I'm not sure why everything looks OK except the tRAS timing.

PC Wizard

General Information :

DIMM0 (BANK0 ) : 2048MB - DIMM

DIMM1 (BANK1 ) : 2048MB - DIMM

DIMM2 (BANK2 ) : 2048MB - DIMM

DIMM3 (BANK3 ) : 2048MB - DIMM

Information SPD EEPROM (DIMM0) :

Manufacturer : Kingston

Part Number : 2G-UDIMM

Serial Number : 353CF094

Type : DDR2-SDRAM PC2-6400 (399MHz) - [DDR2-800]

Format : Regular UDIMM (133.35 x 3)

Size : 2048MB (2 ranks, 8 banks)

Module Buffered : No

Module Registered : No

Module SLi Ready (EPP) : No

Width : 64-bit

Error Correction Capability ... No

Max. Burst Length : 8

Refresh : Reduced (.5x)7.8, Self Refresh?s

Voltage : SSTL 1.8v

Prefetch Buffer : 4-bit

Manufacture : Week 11 of 2011

Supported Frequencies : 200MHz, 266MHz, 400MHz

CAS Latency (tCL) : 3 clocks @ 200MHz, 4 clocks @ 266MHz, 5 clocks @ 400MHz

RAS to CAS (tRCD) : 3 clocks @ 200MHz, 4 clocks @ 266MHz, 5 clocks @ 400MHz

RAS Precharge (tRP) : 3 clocks @ 200MHz, 4 clocks @ 266MHz, 5 clocks @ 400MHz

Cycle Time (tRAS) : 9 clocks @ 200MHz, 12 clocks @ 266MHz, 18 clocks @ 400MHz

Min TRC : 12 clocks @ 200MHz, 16 clocks @ 266MHz, 23 clocks @ 400MHz

[Kryton]

If you are basing this on the CPU-Z image in your post, you are looking at the wrong tab for your actual settings and speeds.

The tab you have shown shows the timings table, a guide for what timings are specified for the speeds in MHz shown above each column.

The memory tab shows your actual/current RAM timings and speed as shown in the screenie I made as an example. My current timings are CL7-8-8-22-32 @ 668.9MHz when the screenie was taken. The SPD tables show how I "should" have it set as shown in the columns based on RAM speed in MHz per specs but as you can see, I have them set differently.

It's also helpful to know not every motherboard manufacturer will run the same brand & model of sticks the same way by default and that's based on how the BIOS itself is setup. In some cases the machine will not run the sticks at the specified specs and a BIOS setting adjustment is needed to set them correctly.

In your case, you have a locked BIOS being it's an OEM machine so not much you can do there regardless.

[Cheryl_27]

Thanks Kryton. In my case, CPU-Z memory & SPD tabs show the same expected & actual timings - even for tRAS (18 clocks instead of very clear specs in Kingston's detailed spec sheet). Maybe spec sheet is wrong, but not likely.

You may be right about the BIOS limiting clock settings - on JUST ONE setting - tRAS?? The BIOS / mobo allowed all other clock settings to run at Kingston's programmed SPDs, (at 800MHz & 1.8v, which is what BIOS setup shows it's recognizing for the Kingston). Maybe HP had a reason to set the BIOS so tRAS wouldn't be lower than 18.

Why limit that one clock setting in particular? It's not like the mobo & all chipsets / specs were low end - exactly opposite, when new.

I'm not an expert on tRAS settings on actual performance, all else being equal (specifically here, w/ 5-5-5). I know what tRAS is / does, but I'd have to research how to calculate it's effect.

If was significant enough, could look into modding the BIOS. Performance diff between CL 5 & 6 is significant (for example).

I do wonder if reported tRAS in CP Wizard / CPU-Z is off - affected by some BIOS setting?

Not necessarily looking to seriously overclock, but I bought extra GBs, w/ faster times for extra performance. I'll run memtest on it anyway - be interesting if it shows diff timing settings.

[articuno1au]

Yeah, that's probably max speeds. You can modify the timings in your bios.

A lot of DDR kits allow timings like 8-8-8-24 (DDR3 @ 1600) but you have to set those timings manually.. Same with voltages..

[Jason S. Global Moderator]

Yeah, that's probably max speeds. You can modify the timings in your bios.

A lot of DDR kits allow timings like 8-8-8-24 (DDR3 @ 1600) but you have to set those timings manually.. Same with voltages..

the OP cant modify the settings b/c it's an HP OEM board.

i had a motherboard a few years ago that forced you to change timings when all 4 DIMM slots were populated. perhaps this is the same situation.

[articuno1au]

Well she's stuck with the 5-5-5-18 then >.<

Kinda sucks :\ You probably won't notice tbh :)

[chconline Veteran]

5-5-5-15... 5-5-5-18... same thing :p

You won't even see a difference in benchmarks to be honest, haha.

[Cheryl_27]

Thanks Gents,

Jdawg683

i had a motherboard a few years ago that forced you to change timings when all 4 DIMM slots were populated. perhaps this is the same situation.

"forced you to," or forced it on you? Did it force changing only ONE timing, like mine & was it a mobo w/ any control over timings?

I know the question "why would HP limit ONLY the tRAS clock" may be a "who the hell knows" type. Someone knows. Think I'll dedicate my life to finding the answer. On 2nd thought...

Total guess, but since tRAS setting is said to be

Active to Precharge Delay (tRAS): After an Active command is issued, another Precharge command cannot be issued until tRAS has elapsed. So, this parameter limits when the memory can start reading (or writing) a different row.

maybe once all DIMM slots are filled, (some) BIOS limit this setting to prevent bottleneck.

5-5-5-15... 5-5-5-18... same thing :p

You won't even see a difference in benchmarks to be honest, haha.

Maybe. Since I've never read details on calculating tRAS' effect on performance. Don't blame anyone for not knowing the answer off top of their head.

Articuno1au & chconline, I assume your comments "probably won't notice the difference" (between 5-5-5-15 & 5-5-5-18) are based on doing / researching benchmarks w/ those timings? Course, this is DDR2 PC 6400, not DDR3.

For grins, can you point me in right direction to get benchmark info for these timing differences, on a system like mine. Seems it'd only be relevant comparing to other Core 2 Quad LGA775 boards w/ similar speed CPU (Q9400). Maybe not. I'm getting into unfamiliar waters here.

Thanks.

[chconline Veteran]

Articuno1au & chconline, I assume your comments "probably won't notice the difference" (between 5-5-5-15 & 5-5-5-18) are based on doing / researching benchmarks w/ those timings? Course, this is DDR2 PC 6400, not DDR3.

For grins, can you point me in right direction to get benchmark info for these timing differences, on a system like mine. Seems it'd only be relevant comparing to other Core 2 Quad LGA775 boards w/ similar speed CPU (Q9400). Maybe not. I'm getting into unfamiliar waters here.

Thanks.

I review RAM.

Obviously I haven't touched DDR2 since 2007/2008, but your situation is pretty much like setting DDR3-1600 at 9-9-9-27 vs 9-9-9-24. It doesn't even make a difference in the benchmarks. Heck, you can run your RAM at 4-4-4-12 and it will make absolutely no difference in real life.

Here's one of my old reviews back in 2007:

http://aphnetworks.com/reviews/a_data_extreme_edition_ddr2_800_2x1gb

I also have a Core 2 Extreme QX9650 and that doesn't make a difference either with regards to RAM.

And if you just look at one of my recent articles, the KHX runs at 9-9-9-27 vs. the Sniper running at 9-9-9-24, any difference is simply negligible.

http://aphnetworks.com/reviews/kingston_hyperx_khx1600c9d3x2k2_8gx_2x4gb

[Jason S. Global Moderator]

i realize you just want to know why the timings are different, but you will never notice the difference. the last number in the sequence, the one that changed from 15 to 18, is also the least important timing and the one that shows little difference in benchmarks. the 1st number, however, is the most important and would show the most difference in benchmarks. either way, you'll see no difference in your daily computing.

[Cheryl_27]

Thanks to all for input.

chconline,

1. Thanks for linking the benchmark tests. Very informative. As you said, similar RAM w/ diff tRAS settings show negligible difference. Unless one's into counting nano sec's. Hey, you waste 1000 ns here & there, & before you know it, the whole day's shot.

2. Some cat named Jonathan Kwan did those reviews.

Jdawg, yes - I have an inquiring mind. A blessing & curse. Sometimes gets the best of me, but also made a lot of $.

I was just as interested in why it lowered tRAS as the performance effect.

[chconline Veteran]

2. Some cat named Jonathan Kwan did those reviews.

Yeah, and that's me.

[TokiToki]

Thanks Kryton. In my case, CPU-Z memory & SPD tabs show the same expected & actual timings - even for tRAS (18 clocks instead of very clear specs in Kingston's detailed spec sheet). Maybe spec sheet is wrong, but not likely.

You may be right about the BIOS limiting clock settings - on JUST ONE setting - tRAS?? The BIOS / mobo allowed all other clock settings to run at Kingston's programmed SPDs, (at 800MHz & 1.8v, which is what BIOS setup shows it's recognizing for the Kingston). Maybe HP had a reason to set the BIOS so tRAS wouldn't be lower than 18.

The first screenshot tells me that Kingston programmed the SPD at 5-5-5-18 and not 5-5-5-15 as according to the spec sheet. Most OEM BIOS will set the timings based on SPD which is why you're getting tRAS 18. Yes, some BIOS do relax timings if all DIMM's are populated, but that isn't the case here.

[Cheryl_27]

The first screenshot tells me that Kingston programmed the SPD at 5-5-5-18 and not 5-5-5-15 as according to the spec sheet. Most OEM BIOS will set the timings based on SPD which is why you're getting tRAS 18. Yes, some BIOS do relax timings if all DIMM's are populated, but that isn't the case here.

Yeah, I noticed that, too. Not exactly sure why it's happening. I may sue Kingston for false advertising & mental anguish.

[Kryton]

I believe it is possible to adjust the timings even if the BIOS is locked.

You can try CPU Tweaker and see if it will work - Here's a link for downloading:

http://www.tweakers.fr/cputweaker.html

If for some reason it doesn't work, maybe Memset will do the trick:

http://www.tweakers.fr/memset.html

Good luck!

[Cheryl_27]

Thanks Kryton. Interesting. I try to always check out software from reliable sources before installing just anything. Maybe that's why I rarely have OS problems.

Not sure CPU Tweaker supports Core 2 Quad CPUs (some said no).

In Mem Tweaker documentation (I found) never mentions supports Intel G33 chipset - EXCEPT in one screen shot showing "Intel P35/G33/G31." Though I may try it, others will probably appreciate this info. Maybe support for "P35" implies G33/ G31 also. Any hardware detection tools (CPU-Z, etc) always show "P35/G33/G31" on my system. Only way I * think * mine's a G33, is mobo specs, or maybe remove chipset heat sink & look.

Developer's site lists an offsite "discussion forums" for their software.

[Kryton]

Thanks Kryton. Interesting. I try to always check out software from reliable sources before installing just anything. Maybe that's why I rarely have OS problems.

Not sure CPU Tweaker supports Core 2 Quad CPUs (some said no).

In Mem Tweaker documentation (I found) never mentions supports Intel G33 chipset - EXCEPT in one screen shot showing "Intel P35/G33/G31." Though I may try it, others will probably appreciate this info. Maybe support for "P35" implies G33/ G31 also. Any hardware detection tools (CPU-Z, etc) always show "P35/G33/G31" on my system. Only way I * think * mine's a G33, is mobo specs, or maybe remove chipset heat sink & look.

Developer's site lists an offsite "discussion forums" for their software.

I believe CPU Tweaker is what you'd need to use.

Memset is definitely for the older systems - May or may not work with your setup.

I have used both and even have these archived back for my OC'ing or whatever else. If it states it's compatable with your chipset, it should work.

Benchmarking utils can be useful even if you're not benching or OC'ing for simply tweaking things the way you want it to be.

[Cheryl_27]

Thanks. Good info. What I'd also really like (along same lines) is utility to unlock this OEM HP / Asus mobo's fan control capability, so 3rd party utilities can control case, if not CPU fans.

Fairly sure it has the ability - hidden / locked. It's very similar in specs / features to some "genuine" Asus boards, but BIOS doesn't offer any fan info or fan control options, as similar Asus boards / BIOSs do. This was quite a feature / expansion rich OEM mobo - custom ordered box from HP's highest end line. Speedfan & others can find / report fan speeds & most temps, but unable to adjust fans (doesn't show options to try).

Read this is probably due to the OEM BIOS, not board capability. Not sure. For what ever reason, the EVGA GPU card fan can be controlled well, by tool like MSI Afterburner, though Speedfan could not.

Seem if prgms like CPU-Tweaker & Memset exist, allowing changing clocks - otherwise impossible w/ existing BIOS, might be a similar prgm to "unlock" fan control.

[Kryton]

Thanks. Good info. What I'd also really like (along same lines) is utility to unlock this OEM HP / Asus mobo's fan control capability, so 3rd party utilities can control case, if not CPU fans.

Fairly sure it has the ability - hidden / locked. It's very similar in specs / features to some "genuine" Asus boards, but BIOS doesn't offer any fan info or fan control options, as similar Asus boards / BIOSs do. This was quite a feature / expansion rich OEM mobo - custom ordered box from HP's highest end line. Speedfan & others can find / report fan speeds & most temps, but unable to adjust fans (doesn't show options to try).

Read this is probably due to the OEM BIOS, not board capability. Not sure. For what ever reason, the EVGA GPU card fan can be controlled well, by tool like MSI Afterburner, though Speedfan could not.

Seem if prgms like CPU-Tweaker & Memset exist, allowing changing clocks - otherwise impossible w/ existing BIOS, might be a similar prgm to "unlock" fan control.

I'd need to look around and see if such a program exists for controlling the case fans themselves but I don't recall ever seeing something like that.

I do know some proggys will allow for CPU fan control, most folks simply buy an aftermarket fan for their case that's rated at a higher CFM. I have a few fans for my CPU's that are variable speed with a speed controller. My V1 CPU cooler is one such item and it does a good job for what it is but it's also a big unit. Luckily I have my setup on a testing stand instead of a case so that's not a problem, case fans are definitley no prob here.