G.Skill Titan 256GB SSD

[VaultDweller]

Originally posted by: TimBob
Already got a response:

Dear customer



Yes Titan SSD should not have a lot of stuttering. You don?t need to tweak this drive. The only thing you need to do is disable the super fetch. If disable super fetch doesn?t help, please rma with us.



Thank you

GSKILL SUPPORT

Please do keep us posted on whether the replacement works any better.

 

[imported_TimBob]

Originally posted by: VaultDweller

Originally posted by: TimBob
Already got a response:

Dear customer



Yes Titan SSD should not have a lot of stuttering. You don?t need to tweak this drive. The only thing you need to do is disable the super fetch. If disable super fetch doesn?t help, please rma with us.



Thank you

GSKILL SUPPORT

Please do keep us posted on whether the replacement works any better.

Haven't started the RMA process yet.

 

[JeffreyLebowski]

Originally posted by: Denithor
Do the happy dance!

This is what I've been waiting for - the makers to fix that annoying stutter - and for prices to come down while capacities go up.

These drives are going to be a lot more useful shortly and within reach of more people.

2009 - Year of the SSD!

Now, are WD/Seagate/etc working on their own models or should I go ahead and short their stocks?

Hold your horses there. SSD's are cool and all, but for the GB per $ ratio, the platter drives are going to be much much better. And when it comes to volume sales, the SSD's just won't be able to compete for at least a year.
Geeks want SSD's for speed, the masses want GB's for storage of family pictures and video and stuff they download.

 

[taltamir]

When asked about their plans in regards to SSD, the CEO of Seagate said that SSD (the technology in general) infringes on their patents.

So... shorting their stock... maybe?

 

[coolVariable]

Originally posted by: JeffreyLebowski
Hold your horses there. SSD's are cool and all, but for the GB per $ ratio, the platter drives are going to be much much better. And when it comes to volume sales, the SSD's just won't be able to compete for at least a year.
Geeks want SSD's for speed, the masses want GB's for storage of family pictures and video and stuff they download.

So you are actually investing for a time horizon sub-1 year?
You have to be kidding me.

Long term (1+ year), SSDs will kill HDDs because the room for improvement could theoretically be the same as for CPUs.

The REAL question is, out of the 100+ companies throwing SSD product on the market now ... who will be around in 5 years? Who will be market leader?

My guess would be: intel, Samsung ... some others.

 

[aka1nas]

There's only 3 or 4 companies actually making the SSDs and their components. Everyone else is just assembling and/or rebadging. No reason we won't continue to have a ton of vendors continue to do so, just like with RAM.

 

[StuartR]

Originally posted by: jedisolo
I own this drive and I have been using it in my Thinkpad T400 since I bought the drive 2 weeks ago, I couldn't have been happier with my purchase.

Count me as another happy owner of two of these Titan drives (128GB). Also, here are my comments regarding the OCZ SSD "tweaks" mentioned here and in their user forum:

After installing my new G.Skill SSD's and re-reading some of the "Tweak" information available on the OCZ forum, I have to take exception to a few of their user recommendations:

Here are a few examples:

One recommendation is a change in "SecondLevelDataCache". This setting is only useful for computers with direct-mapped L2 caches. I understand newer computers (Pentium II and later) don't have direct-mapped L2 caches.

Also, the recommendation to enable "clearPageFileAtShutdown" doesn't seem correct (especially since another recommendation is to remove a pagefile if you have at least 4GB of RAM). Enabling this can actually degrade performance and was only designed as a security feature.

Attempting to enable advanced write caching performance on an SSD doesn't seem appropriate either (and it may not even stay enabled across subsequent boots since the disk controller doesn't have a battery-backed up write cache), although laptops containing a charged battery may prove an exception.

On the other hand, disabling the default defragmenter does seem appropriate as its use could defeat the wear-leveling file positioning intentionally done by the SSD firmware.

Other "tweaks" such as disabling Prefetch, Superfetch and Indexing may or may not noticebly improve performance, depending on the software load you present to the OS.

NCQ in Solid-State Drives
----------------------------
The OCZ SSD installation notes recommend not using an AHCI SATA driver with their SSD. However there may be benefits from one of the AHCI features "Native Command Queuing" (NCQ). NCQ is used in newer solid-state drives where the drive encounters latency on the host, rather than the other way around. For example, Intel's X25-E Extreme solid-state drive uses NCQ to ensure that the drive has commands to process while the host system is busy processing CPU tasks.

So given all this information, my recommendation is to take some of these "tweaks" with a grain of salt. Try them (one at a time) and restore their default settings if you can't see any difference in performance. The key here is to make only one change at a time (and keep a log of what you did and when).

 

[Idontcare]

Originally posted by: StuartR
On the other hand, disabling the default defragmenter does seem appropriate as its use could defeat the wear-leveling file positioning intentionally done by the SSD firmware.

Defragmenting an SSD seems like an entirely worthless endeavor. I had not really thought about it before, but yeah these SSD's ought to come with a windows utility to make it snap-easy for a user to install SSD, install utility/driver, and automatically have stupid stuff like windows defrag disabled on the SSD.

I guess Diskeeper is going to be out of business in about 5 yrs once the spindle business dies.

 

[StuartR]

Originally posted by: Idontcare

Originally posted by: StuartR
On the other hand, disabling the default defragmenter does seem appropriate as its use could defeat the wear-leveling file positioning intentionally done by the SSD firmware.

Defragmenting an SSD seems like an entirely worthless endeavor. I had not really thought about it before, but yeah these SSD's ought to come with a windows utility to make it snap-easy for a user to install SSD, install utility/driver, and automatically have stupid stuff like windows defrag disabled on the SSD.

I guess Diskeeper is going to be out of business in about 5 yrs once the spindle business dies.

Diskeeper is struggling with support for SSD's now. Their Diskeeper V9 Professional product attempts to mitigate SSD fragmentation. I took a cursory look at the product but then uninstalled it, unconvinced so far that it's that useful.

 

[allthatisman]

Originally posted by: StuartR

Originally posted by: jedisolo
I own this drive and I have been using it in my Thinkpad T400 since I bought the drive 2 weeks ago, I couldn't have been happier with my purchase.

Count me as another happy owner of two of these Titan drives (128GB). Also, here are my comments regarding the OCZ SSD "tweaks" mentioned here and in their user forum:

After installing my new G.Skill SSD's and re-reading some of the "Tweak" information available on the OCZ forum, I have to take exception to a few of their user recommendations:

Here are a few examples:

One recommendation is a change in "SecondLevelDataCache". This setting is only useful for computers with direct-mapped L2 caches. I understand newer computers (Pentium II and later) don't have direct-mapped L2 caches.

Also, the recommendation to enable "clearPageFileAtShutdown" doesn't seem correct (especially since another recommendation is to remove a pagefile if you have at least 4GB of RAM). Enabling this can actually degrade performance and was only designed as a security feature.

Attempting to enable advanced write caching performance on an SSD doesn't seem appropriate either (and it may not even stay enabled across subsequent boots since the disk controller doesn't have a battery-backed up write cache).

On the other hand, disabling the default defragmenter does seem appropriate as its use could defeat the wear-leveling file positioning intentionally done by the SSD firmware.

Other "tweaks" such as disabling Prefetch, Superfetch and Indexing may or may not noticebly improve performance, depending on the software load you present to the OS.

NCQ in Solid-State Drives
----------------------------
The OCZ SSD installation notes recommend not using an AHCI SATA driver with their SSD. However there may be benefits from one of the AHCI features "Native Command Queuing" (NCQ). NCQ is used in newer solid-state drives where the drive encounters latency on the host, rather than the other way around. For example, Intel's X25-E Extreme solid-state drive uses NCQ to ensure that the drive has commands to process while the host system is busy processing CPU tasks.

So given all this information, my recommendation is to take some of these "tweaks" with a grain of salt. Try them (one at a time) and restore their default settings if you can't see any difference in performance. The key here is to make only one change at a time (and keep a log of what you did and when).

I would love to see some benchmarks of your RAID0 (I'm assuming) Titan 128gb setup. I am going to probably buy another once the price goes down to sub$300 level, and once Newegg gets them back in. Preferably an HDTune screenshot... if you want to that is

 

[taltamir]

Originally posted by: aka1nas
There's only 3 or 4 companies actually making the SSDs and their components. Everyone else is just assembling and/or rebadging. No reason we won't continue to have a ton of vendors continue to do so, just like with RAM.

what confuses me is, how come several start ups as well as all the ram manufacturers are rebadging and selling SSDs, yet not a single HDD manufacturer has chosen to do the same...

It stinks of large corporations trying to pretend a new, better, product does not exist in the hope it will fail (it never does) instead of demolishing their business (it always does). Maybe they are worried that if they started selling them people will find out about them, and eventually go and buy it from the source... from companies like intel and samsung and jmicron.

 

[Idontcare]

Originally posted by: taltamir

Originally posted by: aka1nas
There's only 3 or 4 companies actually making the SSDs and their components. Everyone else is just assembling and/or rebadging. No reason we won't continue to have a ton of vendors continue to do so, just like with RAM.

what confuses me is, how come several start ups as well as all the ram manufacturers are rebadging and selling SSDs, yet not a single HDD manufacturer has chosen to do the same...

It stinks of large corporations trying to pretend a new, better, product does not exist in the hope it will fail (it never does) instead of demolishing their business (it always does). Maybe they are worried that if they started selling them people will find out about them, and eventually go and buy it from the source... from companies like intel and samsung and jmicron.

What you are witnessing is the concept of disruptive technology in action.

http://en.wikipedia.org/wiki/Disruptive_technology

SSD is a disruptive technology relative to the established (and mature) industry built around spindle-based hard-drive technology.

By nature of it being a mature industry, the current crop of leaders tend to have personalities that prevent them from being risk-takers.

How many car manufacturers do you know of that were originally horse-buggy manufacturers?

In 5 yrs time all these little startups will either have long since been bought out by the Seagates and the Hitachis or they will have replaced them as the industry titans.

It will be an interesting world in their new business model. Supply-chain wise the equivalent would be if current hard-drive makers did not do their own platter manufacturing or research. Imagine if Seagate and Hitachi bought platters sold wholesale on an open market, then they assembled the drives and for product differentiation the magic sauce was added in at the PCB level (jmicron vs. samsung vs Intel, etc) and the drive motor mechanics. Things like areal density and durability would be determined by a whole other entity in the supply chain (the platter manufacturers).

That is what business will be like in five years, you have the chip makers who pump out flash chips that meet Jedec standards which any SSD maker can buy and interchange with their device (lowering costs by leveraging industry volumes for the chips). It's great for everyone except the shareholders and employees of the traditional, about to be extincted, platter drive manufacturers.

 

[taltamir]

I think it has to do with the publically traded corporation... there is no single individual who says "this is MY company, MY money, and I am not letting it go down". Instead there are a variety of employees who are pulling a check, if the company goes down, they will move on to another one. They can't risk their job, and it is going to take a LOT of bashing your head against the wall to move the company anywhere, and if you DO you are not rewarded for it... So nobody does it.

In 5 yrs time all these little startups will either have long since been bought out by the Seagates and the Hitachis or they will have replaced them as the industry titans.

Only if those startups are foolish. They have nothing to gain from such a buy. Netflix is pushing blockbuster to bankrupcy... same will happen here.

It will be an interesting world in their new business model. Supply-chain wise the equivalent would be if current hard-drive makers did not do their own platter manufacturing or research. Imagine if Seagate and Hitachi bought platters sold wholesale on an open market, then they assembled the drives and for product differentiation the magic sauce was added in at the PCB level (jmicron vs. samsung vs Intel, etc) and the drive motor mechanics. Things like areal density and durability would be determined by a whole other entity in the supply chain (the platter manufacturers).

Not for everyone... intel for example makes their own, and it shows. I remember there being a startup or two who entered the SSD market YEARS ago who ALSO make their own...

As for disruptive tech... very interesting... flash is both a new market (ultra durable) and a low end (speed, size, cost). Or... it was... it is now faster or slower depending on the load for top end drives... size and cost need to improve but will..
But yea, 5 years from now the incumbents will be extinct... OR they will have purchased some of the challengers for a huge loss...

 

[taltamir]

hi, so since I am hearing conflicted reports about the titan stuttering... basically I am concluding that it stutters, but LESS... so how much less? how much better is it?

 

[aka1nas]

If you apply most of the basic SSD tweaks (move page files and problem apps off the SSD), than the drive almost never stutters during normal usage. It happens maybe 3-4 times a week for me now. I have my OS and all of my games running off the SSD, and I can pretty comfortably install an app or two to the SSD and use the PC without anything locking up. It's held up well to gaming, too.

Problem apps for the SSD include:

Filesharing apps
Firefox (tons of constant random writes to the profile and cache folder, having these on the SSD made it stutter a lot)

Basically, anything that constantly writes small pieces of data to disk will perform poorly on the SSD. Not really an issue if you have some mechanical disks still to offload things to. If you want something that's completely plug-and-play, wait for something like the Vertex or the new samsungs if they review well.

Based on the experience people are having with SSDs on hardware RAID controllers with dedicated write cache, I think we will eventually see all SSDs come with some amount of onboard cache to alleviate potential stuttering.

 

[Idontcare]

Originally posted by: taltamir
Not for everyone... intel for example makes their own, and it shows. I remember there being a startup or two who entered the SSD market YEARS ago who ALSO make their own...

There's always the exception to the rule that can be dragged out. Micron and Crucial were no different for DRAM and they will be no different for SSD's.

The flash Intel uses in their SSD's are actually not produced directly by Intel but rather by a company called IM Flash which was jointly formed between Intel and Micron (ergo IM...Flash). Just like with Crucial, it will only be a matter of time before the Intel SSD division is allowed to buy Flash chips from whoever will sell them the cheapest on the open market. If Samsung undercuts IM Flash prices for Intel SSD's there will be no loyalty to IM Flash at some point.

All these joint ventures devolve to this, they have to because Intel management is beholden to Intel shareholders and not IM Flash shareholders. But we are digressing waaaay off-topic here.

 

[Tweakin]

Originally posted by: Old Hippie
Have you tried the little SSD Tweaker?

I'm not getting the stuttering issues but it seems to make everything a little faster.

After looking at the tweaks...why wouldn't these also work for hardware based storage?

 

[Idontcare]

Originally posted by: Tweakin

Originally posted by: Old Hippie
Have you tried the little SSD Tweaker?

I'm not getting the stuttering issues but it seems to make everything a little faster.

After looking at the tweaks...why wouldn't these also work for hardware based storage?

These tweaks will work with non-SSD drives, but you lose performance.

The things that people are undoing with these tweaks were originally conceived and implemented as a way to mitigate some performance limiting aspects of spindle-based drives.

It just turned out much later that these same tweaks work against SSD drives in a way that reduces their performance. So undoing the tweaks helps with SSD's, and while they won't break your non-SSD drives you will be undoing things that weren't intentionally implemented to make them faster so expect some things to be slower.

 

[Elfear]

Would it be worth running a pair of the Titan SSDs (or any newer SSD) in RAID 0 with the onboard RAID controller? Dedicated RAID controllers are a little pricey for a good one from what I've read.

 

[aka1nas]

Originally posted by: Elfear
Would it be worth running a pair of the Titan SSDs (or any newer SSD) in RAID 0 with the onboard RAID controller? Dedicated RAID controllers are a little pricey for a good one from what I've read.

Normally, you would be looking for a controller with dedicated write cache to eliminate the stuttering. However, I have read some reports on XS about people getting excellent results with ICH10R.

If you go with a dedicated card, there are a few that are "merely" expensive that still have dedicated cache memory.

http://www.newegg.com/Product/...x?Item=N82E16816115047

http://www.newegg.com/Product/...x?Item=N82E16816116047

 

[Elfear]

Originally posted by: aka1nas

Originally posted by: Elfear
Would it be worth running a pair of the Titan SSDs (or any newer SSD) in RAID 0 with the onboard RAID controller? Dedicated RAID controllers are a little pricey for a good one from what I've read.

Normally, you would be looking for a controller with dedicated write cache to eliminate the stuttering. However, I have read some reports on XS about people getting excellent results with ICH10R.

If you go with a dedicated card, there are a few that are "merely" expensive that still have dedicated cache memory.

http://www.newegg.com/Product/...x?Item=N82E16816115047

http://www.newegg.com/Product/...x?Item=N82E16816116047

Thanks for the advice aka1nas. In your opinion, is it better to go with a pair of SSDs with the new controllers using onboard RAID or buy a dedicated RAID controller and use some of the older SSDs? Price wise it would be about the same.

 

[StuartR]

No Raid 0 here. The SSD's are in an HP HDX Dragon "laptop" which has two drives.

 

[aka1nas]

The dual controller drives like the Titan and Apex would mostly be preferable due to their higher transfer rates. I would get a second opinion on the onboard RAID (maybe from XS?) before committing to that setup. I think one guy on their Titan thread is running two 128s with ICH10R and not experiencing any stutters.

 

[genec57]

I have read lots of chatter here about G.Skill Titans and stutter and other evils. I agree with one previous poster that most of the criticism seems to be from those who have opinions but no SSD experience.
I have two GSkill Titan 128s in RAID 0 and they are quite simply smokin fast. Sequential reads average at 350mb/s. I think (not sure) there may have been some stutter early on in that my mouse would hesitate in Firefox. I moved the Firefox cache to my Data array and the problem has not recurred.
You want to be sure that drive caching is enabled - that's important. Superfetch, prefetch seems to make no difference one way or the other so I leave them on. Defrag definitely off.
I have had the array up and running continuously for almost three weeks now and there has been NO degradation of performance and I have installed and reinstalled multiple apps.
The absolute only negative I can find is cost.

 

[genec57]

double post