How to set SATAII in Windows XP?

[Senator]

I have two WD HDs that are SATAII w/16MB Cache.

However, Intel desktop Utilities Version 2.1 Light is saying
Max transfer mode = UDMA 6(ATA/133)
Active Transfer Mode = UDMA 5 (ATA/100)

They don't seem to be transferring as fast as possible.

What do I need to do to optimize these drives? Thanks!

 

[RichPLS]

What are your BIOS settings for those drives? Do you have the SATA ports set to mimic IDE emulation?

 

[The_Prophecy]

I have two WD HDs that are SATAII w/16MB Cache.

However, Intel desktop Utilities Version 2.1 Light is saying
Max transfer mode = UDMA 6(ATA/133)
Active Transfer Mode = UDMA 5 (ATA/100)

They don't seem to be transferring as fast as possible.

What do I need to do to optimize these drives? Thanks!

You don't have to do anything. They are already either transferring at 150MB/s if you have them set to run on a SATA1 bus, or at 300MB/s if it's running on a SATA2 bus... the BIOS and all programs that report these specs just read the active transfer mode as UDMA 5 instead of (what should technically be) UDMA 8 (UDMA 7 being the SATA1 spec of 150MB/s and UDMA 8 being the SATA2 spec of 300MB/s). This is perfectly normal, you do not have to do anything.

 

[linux_0]

I have two WD HDs that are SATAII w/16MB Cache.

However, Intel desktop Utilities Version 2.1 Light is saying
Max transfer mode = UDMA 6(ATA/133)
Active Transfer Mode = UDMA 5 (ATA/100)

They don't seem to be transferring as fast as possible.

What do I need to do to optimize these drives? Thanks!

You don't have to do anything. They are already either transferring at 150MB/s if you have them set to run on a SATA1 bus, or at 300MB/s if it's running on a SATA2 bus... the BIOS and all programs that report these specs just read the active transfer mode as UDMA 5 instead of (what should technically be) UDMA 8 (UDMA 7 being the SATA1 spec of 150MB/s and UDMA 8 being the SATA2 spec of 300MB/s). This is perfectly normal, you do not have to do anything.


Indeed. :-D

You are not going to get 300MB/sec btw.

The fastest SCSI drive only gets 97MB/sec on a U320 Interface.

300MS/sec is the interface speed not the drive speed.

 

[The_Prophecy]

You're right... that is the interface speed, however the basic idea behind my post was that the BIOS doesn't read that properly and he has nothing to worry about.

 

[linux_0]

You're right... that is the interface speed, however the basic idea behind my post was that the BIOS doesn't read that properly and he has nothing to worry about.

:-D

Indeed the controller and drives will usually negotiate the best mode they can agree on.

The BIOS sometimes provides incorrect information. The Intel software may also be providing incorrect information.

 

[pat]

Just keep in mind that SATA is only an interface, and that the controller and drive are still IDE devices, they still use the same ATA command, but SATA add a few more than the normal ATA one.

 

[mesarectifier]

Errr, SATA2 doesn't exist yet, does it? Do you not just mean SATA 300? Or is it I who is the confused one.

 

[The_Prophecy]

SATA2, SATAII, SATA 300, SATA 3Gbs <---- all interchangable.... mean the exact same thing. And SATA2 has been out for 6 months - 1 year now...

 

[linux_0]

SATA2, SATAII, SATA 300, SATA 3Gbs <---- all interchangable.... mean the exact same thing. And SATA2 has been out for 6 months - 1 year now...

Indeed, but it's kind of a mess because people were not using the name consistently.

With the release of the NVIDIA nForce4 chipset in 2004, the clock rate was doubled to 3.0 GHz, for a maximum throughput of 300 MB/s. SATA 3Gb/s is backward compatible with SATA 1.2Gb/s, allowing SATA 1.2Gb/s hardware to interface with SATA 3Gb/s ports and vice versa. However, some systems that do not support SATA speed autonegotiation may require that the drive's speed be manually limited to 150 MB/s with the use of a jumper for a 300 MB/s drive.[1]

The 3.0 GHz specification has been very widely referred to as “Serial ATA II” (“SATA II”), contrary to the wishes of the Serial ATA standards organization that authored it. The official website notes that SATA II was in fact that organization's name at the time, the SATA 3Gb/s specification being only one of many that the former SATA II defined, and suggests that “SATA 3Gb/s” be used instead. (The Serial ATA standards organization has since changed names, and is now “The Serial ATA International Organization”, abbreviated SATA-IO.)

SATA-IO plans to further increase the maximum throughput of Serial ATA to 600 MB/s around the year 2007.

SATA 3Gb/s is also sometimes referred to as SATA/300, continuing the line of PATA/100, PATA/133 and SATA/150.

 

[Codesmith]

No SATAII and SATA 3.0 are not interchangeable

There are currently no different levels or generations of SATA drives PERIOD>

All drives out are the same generation, all implement various features from the SATA I to SATA 2.5 standards.

The interface speed is just one of many features and choice of interface speed
1) has no performance benifit (as shown by Anandtech benchmarks)
2) has nothing to do with what other features the drives have.

That is why the SATA-IO group forbids manufactures from calling their products SATA II.

 

[The_Prophecy]

where do you see SATA 3.0 in my last post? cause i don't.......

 

[linux_0]

No SATAII and SATA 3.0 are not interchangeable

There are currently no different levels or generations of SATA drives PERIOD>

All drives out are the same generation, all implement various features from the SATA I to SATA 2.5 standards.

The interface speed is just one of many features and choice of interface speed
1) has no performance benifit (as shown by Anandtech benchmarks)
2) has nothing to do with what other features the drives have.

That is why the SATA-IO group forbids manufactures from calling their products SATA II.

Wait a minute. I'm not sure I understand what you meant.

I think there is a lot of confusion here.

The SATA standards are in theory backwards and forwards compatible.

The industry and the community have not been using the definitions of the SATA standards the way the SATA II / SATA-IO organization intended

The theoretical interface bandwidth has virtually 0 ( zero ) impact on performance.

 

[Codesmith]

Let me clarify.

1 SATA II is the just the name of a standard and not a label which you are allowed to apply to any real product.

2 The SATA-IO group (who gets to decide these things) doesn't want products to be labeled SATA II because it causes people to think that there are currently two generations of SATA drives or two levels of SATA compliance when that is is not the case.

3 Manufactures like to use the label SATA II for the exact same reasons that the SATA-IO group forbits it.

4 Drives and controllers supporting SATA 150 MBps and SATA 3.0 Gbps are of the same generation and having one tranfer rate or the other says nothing about what other optional SATA features the drive possesses. Instead of two distince levels its just a grab bag where the manufacturers pick an chose what features to throw in a particular SATA drive.

5 Professional benchmarking has so far concluded that SATA 3.0 Gbps vs SATA 150 MBps currently has no effect on performance.

http://www.anandtech.com/storage/showdoc.aspx?i=2760&p=5

We did not notice a performance delta in the 7200rpm drives when enabling 3Gbps operation except in our synthetic benchmarks that measure and report burst speeds.

---

where do you see SATA 3.0 in my last post? cause i don't.......

SATA2, SATAII, SATA 300, SATA 3Gbs <---- all interchangable.... mean the exact same thing. And SATA2 has been out for 6 months - 1 year now...

6 SATA 3.0 Gbps is the correct name, but it is also sometimes called SATA 300 MBps or just SATA 300 because the interface acutally only supports 300 MBps and not 384 MBps inplied by the name 3.0 Gbps.

The SATA-IO group decided that SATA 3.0 Gps and SATA 6.0 Gbps sounded nice. So while the 150 MBps rating is excludes any parity bits the 3.0 Gbps counts those bits padding the result by 20%. So its not a lie so much as an inconsistency in naming conventions.

7 SATA 150 and SATA 3.0 are both part of the same standard which supports both speeds.

8 SATA 3.0 Gbps support is usefull in a controller because there are various storage solutions which connect via a single SATA channel but are capbible of speeds exceeding 150 MBps, for exampe I will be suprised if Gigabyte doesn't release a 3.0 Gbps version of its iRAM product.

Enough for now, CLRMAME has just finished auditing my ROMS and I see that I finally have a 100% complete collection [/quote]

 

[Senator]

thank you. you answered my question while being clear and concise! I guess it's doing the best it can!

Also, I thought we were talking about 3GB/s and not 300MB/s

I have two WD HDs that are SATAII w/16MB Cache.

However, Intel desktop Utilities Version 2.1 Light is saying
Max transfer mode = UDMA 6(ATA/133)
Active Transfer Mode = UDMA 5 (ATA/100)

They don't seem to be transferring as fast as possible.

What do I need to do to optimize these drives? Thanks!

You don't have to do anything. They are already either transferring at 150MB/s if you have them set to run on a SATA1 bus, or at 300MB/s if it's running on a SATA2 bus... the BIOS and all programs that report these specs just read the active transfer mode as UDMA 5 instead of (what should technically be) UDMA 8 (UDMA 7 being the SATA1 spec of 150MB/s and UDMA 8 being the SATA2 spec of 300MB/s). This is perfectly normal, you do not have to do anything.

 

[Codesmith]

No more than ATA 100 vs ATA 133 and having an ATA 100 cable vs an ATA 133 cable.

But its nice for the technology to have room to expand. They already have 6.0 Gbps standard waiting as well.

Plus there are solid state drives that use DDR memory & a battery backup to act like a very fast hard drive, I can see those easily exceeding 150 MBps.

Also you could have an external RAID 5 storage unit connected via a 3.0 Gbps SATA controller. I could see something like that exceeding 150 MBps as well.

Also you never know, maybe the hard drive manufacturers will come up with some trick to break the 150 MBps barrier like adding extra read heads so that a track can be read every half rotation?

 

[linux_0]

thank you. you answered my question while being clear and concise! I guess it's doing the best it can!

Also, I thought we were talking about 3GB/s and not 300MB/s
I have two WD HDs that are SATAII w/16MB Cache.

However, Intel desktop Utilities Version 2.1 Light is saying
Max transfer mode = UDMA 6(ATA/133)
Active Transfer Mode = UDMA 5 (ATA/100)

They don't seem to be transferring as fast as possible.

What do I need to do to optimize these drives? Thanks!

You don't have to do anything. They are already either transferring at 150MB/s if you have them set to run on a SATA1 bus, or at 300MB/s if it's running on a SATA2 bus... the BIOS and all programs that report these specs just read the active transfer mode as UDMA 5 instead of (what should technically be) UDMA 8 (UDMA 7 being the SATA1 spec of 150MB/s and UDMA 8 being the SATA2 spec of 300MB/s). This is perfectly normal, you do not have to do anything.


As Codesmith mentioned the interface speed is either 150MB/sec ( about 1.5Gbps ) or 300MB/sec ( about 3.0Gbps ) with 600MB/sec ( about 6.0Gbps ) due out in the not too distant future. This is Gigabits per second not GigaBytes per second.

SATA uses 8B/10B encoding where 8bits get encoded as 10bits and due to that and other overhead your actual performance is less than you might initially think it is.

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

http://www.anandtech.com/storage/showdoc.aspx?i=2450

Take everything with a grain of salt.

 

[Codesmith]

I never realized it was actually one bit wide at 1.5 or 3.0 Ghz.

1.5 Ghz x .8 (2 parity bits) *1024 (G to M) / 8 (b to B) = 153.6 MBps
3.0 Ghz x .8 *1024/8 = 307.2 MBps

Except the G in a frequency measurement is one billion and while the G in a Gb is (1024^3) as it is in a Byte measurement.

B and b is confusing enough but there is also M vsM and G vs G when comparing freguencies to Bytes. Unless you are selling hard drives and then G in GB suddenly becomes 1000^3 again.

So you need to adjust by (1000/1024)^3 = .93% for the G vs G conversion.

So really is its really 143.05 MBps & 286 MBps?

Just like the PCI bus is actually 127 MBps instead of the often quoted 133 MBps?

(sigh)

Its nearly impossible to find trustworthy statistics because people use 1000 base and 1024 base prefixes interchangibly, often without noticing.

PS yes I know there is "i" notation. So people could say things like 133 MBps = 127MiBps which means you can move 127MiB file in one second.

But no one useds the "i" notation.
There are also conventions about when M is a M and when its an Mi, but these are not consistently followed either.

You OS reports bandwidth in MiBps you ISP in Mbps. CDs and RAM are measured in MiBs, DVDs and Hard drives.

 

[linux_0]

Aye it's a mess.

It's somewhat less of a mess in the Unix world. Some organizations and groups are using MiB and GiB notation already.

 

[Codesmith]

Microsoft needs to addopt the "i" notation for the next version of windows. They have enough market share that it would influcence the community.