Hey guys, becouse of all the problems ive had with this laptop i completely forgot to test the SSD i got for it and someone else reminded me to do that today.
The laptop does feel faster than with the regular 7200rpm HD (boots up in 17 seconds) so i didnt bother testing before, but the benchmarks are telling their horrible little story as you can see.
I tried all the windows provided AHCI drivers and the Intel Rapid Storage one and the difference between those is marginal. I also trimmed the drive before doing the tests.
A reference review to see how the tests should look like can be found HERE.
The screenshots:
![]()
![]()
![]()
![]()
The numbers are simply much lower than they should be (the important random read ones).
Anyone got any idea what i might be missing - is there any solution here?
Or have i perhaps got a bad drive?
Or is it possible these laptops have some cheap controllers that cant handle it?
If anyone has any idea, plese post them.
And if you have one of these SSDs can you please run the same tests on yours and post them so i can see if its the laptops fault?
I know there are some CDM benchies here, but id really like someone posting some more of those and HD Tune ones and if you add your boot up time from the moment you press the power button to when you see the desktop load that would also be great (i know theres a lot of variables at boot up times but im just curious).
-
You mentioned before that you tested the ssd in another laptop, and had full speed. What chipset does that one have? Is it the same HM55?
-
I tested it in the D900F. That one has a desktop X58 chipset. This SSD ran 251mb/s through the entire HD Tune test and got similar results than in the review i linked.
I doubt the SSD is faulty, but i cant absolutely exclude it without testing it again in another machine and at the moment i dont have any with me.
Im more prone to think its the G73JHs fault that the numbers on 4k reads/writes in CDM are so low since in the thread i linked people were getting similar results... But i cant believe im getting such bad results on HD Tune too...
Cheap hardware again? I mean we have two hard drive bays and dont even have raid, god knows what hardware they chose to save money...
And does anyone know what the 4k QD32 is and where is that utilised (at what operations)? Havent heard of that before...
And can anyone do the benchmark set i did? (HDTune is set on default 64k btw) -
Seems like a lot of notbooks have this problem.
One example Link
Does missing 25mb/s on 4k really make a huge difference preformace wise?
I have the C300 128gb and will test it when I get home if you would like. -
Do test it please so i can see what youll get and ill compare it to some decent review to see if its really the G73JH hindering the SSDs performance.
-
I think the issue is coming from the HM55 chipset. This set has no RAID, and may be lacking some advanced features the X58 supports.
-
There are other laptops with mobile i7s which can squeeze the full performance of such a SSD.
But im here to find the cause for our lack of performance. -
Is your bios set to AHCI? If its in IDE mode, thats definately your bottleneck
-
I agree with you Hellcry, SSD performance is a wee bit slow on the G73JH I think. Okay I have a different SSD (Sandforce) but as you can see from the screenies my benches are on the slow side too and I don't know why either. AHCI is enabled (in BIOS and in Windows) and the controller is running with the latest Intel RST driver (9.6.0.1014 AFAIK).
Its a pitty we don't have the same SSD Hellcry but well, this will give us something to compare with at least. Maybe there will be others replying to your post.
All the best
iemie
*EDIT My one boots in 30 seconds. I use this little thingie to findout your boot time Restart Time - Windows 7 Forums -
I have a x25 160 too. my bench are similar as you.
But keep in mind that the bench you saw online are done on a completely empty drive. They were not the boot drive and this impact performance a lot.
-
Here is mine, and it looks like they are just as slow and my is the C300 128GB SATA III
Attached Files:
-
-
Could it be that it needs a firmware upgrade? I'm just reading through sites as i plan on getting the X25 80gb soon. But now i'm not sure about it...
-
its still crazy fast and worth it. I dont notice any difference with my other ssd's that have a faster 4k
-
Got the same capped at 160MB/s results with my Supertalent SSD. The SSD reached some 230MB/s on my desktop where is was builtin before.
I think the Intel Mobile chipset is capped to S-ATA1 speeds, something also Apple does with some of it's powerbooks to save to energy.
It might also be a different issue, but my guess is limitation on the controller side to save energy.
pato -
i safe mode it goes to normal. there is something in the g73 tht slows it down. you wont notice it. just enjoy your ssd
-
Ok, thank you guyz, judging from these benchmarks its definetely the G73JH and whatever controller they used or perhaps some software/driver conflict again. There are laptops with the same mobile chipset which can reach 30mb/s reads and 50mb/s 4k writes so its possible on this chipset too, it just depends on the quality of hardware or software used...
Im a bit dissapointed with this, but i guess you get what you pay for - they saved money on things here and there as we know...
The SSD is still a big improvement over the regular HD, dont get me wrong there, but i do miss my D900F again in this department...
And a note - using the latest Intel Rapid Storage driver didnt yeald the best results, i got a few MB/s more than what i posted up in my original post from the Generation 5 AHCI driver that windows already has and can be chosen at "manual driver update". This one divides into 4 channels and seams to handle it a bit better for some reason. -
Cheers mate, I'll give that driver a try and see how the Sandforce likes that one. I'll keep you posted.
-
HellCry,
Can you clarify how to "manually update" to the Microsoft AHCI Driver? That's exactly the one I've been meaning to use, but after a clean install it does not automatically use this one and my WEI rating only lists 5.9. I've been updating to the intel rst driver to bring that rating up to 7.7, but am very interested in using the Microsoft one instead. Appreciate your help! I'll see if I can run some benchmarks on my x25-m 80gb too. -
Go into device manager, find the AHCI driver, right click and select update driver, select browse my computer for driver software, than let me pick from a list of device drivers on my computer and than select the Generation 5 one there. Let it install and reboot.
-
Ok cool thanks. I'll give it a shot!
-
PulsatingQuasar Notebook Consultant
Maybe you should bring this issue under the attention of Gary. It could be that a BIOS fix is needed because the drivers will be notebook independent. As in, I highly doubt our drivers are the problem.
-
I believe its the onboard controllers fault, but sure ill notify Gary to see what he thinks.
-
Its weird tho that in safe mode we have the correct speed...
I will try to maybe find out what is the culprit of this. -
Some lenovo users had a problem with some power settings - with PCI throttling enabled they got bad results, but with it disabled it ran at full speed. But i have PCI power link managment off... Or is there some other setting i could tweak? I cant think of anything else right now. I mean i use the same settings and tweaks when i install the OS and everything else on all my laptops and only this one has problems with SSD speeds.
Anyway youre welcome to contribute. Let me know if you find something. -
Did you test it in safe mode? I have normal result.
Ive tried killing all service in normal mode and all process to make it "look" like im in safe mode but still can seem to make the speed comeback... -
Good Day,
I have put this on the test list here. It will be a few days before I have results. We will test with SandForce, Indilinx, and Intel drives plus the Crucial if I can get it in time from CA.
As background information, the performance of the storage subsystem is greatly affected by power management settings and of course the chipset. The HM55 is not as fast as the ICH10R when it comes to pure throughput or latency in most cases. Even so, they are fairly close so that leave power management settings as the obvious culprit.
Sincerely,
Gary -
I also get pretty normal results in safe mode (4k 19,4mb/s read and 48,8mb/s write)...
Has to be something that is loaded and runs with windows in normal mode...
Would be nice if you could figure it out Gary, becouse it seams the hardware is capable of doing nearly full speed and it would be a waste to have it almost halved by something here, since we all know just how expensive these drives are and how well they can and are expected perform. -
I have the X25-M G2 80GB version as my system drive, so not quite the same model as you but I would expect similar (if not identical) performance. I didn't get screenshots, but the results I got from CrystalDiskMark 3.0 (for 1000MB and 100MB runs) were as follows:
Code:----------------------------------------------------------------------- CrystalDiskMark 3.0 x64 (C) 2007-2010 hiyohiyo Crystal Dew World : http://crystalmark.info/ ----------------------------------------------------------------------- * MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s] Sequential Read : 254.293 MB/s Sequential Write : 81.172 MB/s Random Read 512KB : 191.187 MB/s Random Write 512KB : 83.760 MB/s Random Read 4KB (QD=1) : 19.731 MB/s [ 4817.0 IOPS] Random Write 4KB (QD=1) : 42.049 MB/s [ 10266.0 IOPS] Random Read 4KB (QD=32) : 152.496 MB/s [ 37230.4 IOPS] Random Write 4KB (QD=32) : 82.816 MB/s [ 20218.7 IOPS] Test : 1000 MB [C: 45.8% (34.1/74.4 GB)] (x5) Date : 2010/08/18 10:39:05 OS : Windows 7 Ultimate Edition [6.1 Build 7600] (x64)
Code:----------------------------------------------------------------------- CrystalDiskMark 3.0 x64 (C) 2007-2010 hiyohiyo Crystal Dew World : http://crystalmark.info/ ----------------------------------------------------------------------- * MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s] Sequential Read : 250.616 MB/s Sequential Write : 84.223 MB/s Random Read 512KB : 196.142 MB/s Random Write 512KB : 84.926 MB/s Random Read 4KB (QD=1) : 18.485 MB/s [ 4513.0 IOPS] Random Write 4KB (QD=1) : 48.028 MB/s [ 11725.5 IOPS] Random Read 4KB (QD=32) : 156.168 MB/s [ 38126.9 IOPS] Random Write 4KB (QD=32) : 84.802 MB/s [ 20703.6 IOPS] Test : 100 MB [C: 46.1% (34.3/74.4 GB)] (x5) Date : 2010/08/18 10:44:28 OS : Windows 7 Ultimate Edition [6.1 Build 7600] (x64)
I don't know what the difference might be, but for what its worth, I'm using the latest Intel 5 series AHCI driver (9.6.0.1014) and do have Link State Power Management set to Off.
Paul -
Will try the above out when I get home
-
OK so I just ran CrystalMark on my intel X25-M G2 80GB SSD and experienced the same exact slow downs that HellCry is seeing with his 160GB unit.
However, booting into Safe Mode made all the results jump up significantly.
I'm currently running with the standard "High Performance" power plan, installed the latest intel RST AHCI driver, and have flashed to BIOS 211. I have a clean install of Windows 7 x64 & do NOT have Power4Gear installed.
PaulH2 what Power Plan are you running?
HellCry, what about you?
Edit: Just adjusted my BIOS overclock setting from TURBO to NORMAL with no impact on the CrystalMark scores. -
where do I get the Intel 5 series AHCI driver (9.6.0.1014) cant seem to find the one I have or this one. Where would it be n the device manager?
-
We can already put chipset performance out of the list of problem since in safe mode, it work correctly.
There is something in "Normal mode" that do something to it.
Maybe ill try reinstalling the x25 and testing it nude with no driver at all except for Intel RST 9.6 -
I don't have an ssd and this might not make sense, maybe someone could try to restart, hit f8 and then select "disable driver enforcement". I don't think it would but its still worth a try since the chip set is able to hit the speeds. So it has to be a software issue.
-
smile_gerard just tried the F8 "disable driver enforcement" suggestion and no luck. Same benchmark result as booting normally. It does not reproduce the elevated performance you see in Safe mode.
-
Good work guys we will find the problem. Will update if I come up with anything.
-
-
I'll go ahead and try throwing Power4Gear on there and see if that helps. I'll report back in a few...
-
Ok so my gut feeling about Power4Gear was accurate... complete bloatware. My CrystalMark scores all dropped compared to the already low default Win7 High Performance Power Plan. I was even using the Power4Gear "High Performance" mode... can only imagine the stellar benchmarks one would get using Quiet Office or Entertainment.
I sure wish there was a way to COMPLETELY disable power management in Windows 7. -
Been doing a search and i came across this.
Outlook, Conclusion And Remaining Questions & Answers : Does Power-Saving Technology Kill SSD Performance?
Look down into the conclusion in particular this sentence
"We found that the sophisticated power saving mechanisms—such as the Active State Power Management for PCI Express, or the deeper C states that switch off entire functional units within the CPU at a transistor level—have a noticeable impact on the performance of our X25-E flash SSD. " -
I already said that lenovo users found that out, but they had software that allowed to completely disable PCI power saving and ive turned off everything that i can in windows advanced power settings and it still isnt enough...
But i used the same settings on my previous laptops and there were no problems, so i cant believe its just the power plan...
Ill wait for Garys response here, im too tired to go reinstalling everything over and over to pinpoint another problem... -
Dell/Alienware PM55 based systems have the same problem.
-
I'll try to run again in safe mode later on just out of interest. -
-
"
12:58 <CiTay> i know that with current intel chipsets and CPUs, they seem to have this "C6 bug" that when they go into a deep power state, the performance of USB/FW/SATA drops
12:59 <CiTay> workaround is to keep CPU busy or disable C6" -
PulsatingQuasar Notebook Consultant
Ehmm.....OK. Disabling it is a BIOS thing right?
Either I am overlooking it or the G73JH BIOS doesn't have that option. -
Ive tried Windows 7 Sp1 beta just for fun and it didnt change a thing for this problem.
-
Ah you had a good idea regarding C6, now that you mentioned it I actually remember to have read it once.
So I fired up for testing the Intelburntest 2.5 and fully loaded my cpu.
And 'oh behold' here the result
It's not perfect (that would be 230-250MB) but around 40-50MB/s better as with low cpu load. -
OMG it work!! under load the speed is super great!
Hellcry can you confirm too?
here is a sexy shot of it in action.
-
This may be an off the mark question, but did anyone perform the install using the F6 and loading the Intel drivers during the install?
Intel X25-M G2 160GB in G73JH - Why the sluggish performance?
Discussion in 'ASUS Gaming Notebook Forum' started by HeavenCry, Aug 17, 2010.