Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
SATA issue after 10.6.3 update
- Thread starter fernando.kreutz
- Start date
- Sort by reaction score
when u hear the startup sound, just press "f" key. it will reload kext cache.
Ok thanks. Is this documented anywhere, I can find any link to this on Google
I think its funny how everyone is wishing that their 1.5 Gbps drives were still falsely showing that they were operating at 3 Gbps, while I would love to be able to force my 3 Gbps drive to operate at 1.5 Gbps, without having to run EFI 1.6! Oh, the irony...
For those that have drives firmware forced down to SATA I speeds, it is now showing the correct speed. Having it show a 3 Gbps rate would make the drive faster, just like this product would make your car go faster:
http://www.youtube.com/watch?v=N8eOTGhTpDY
For those that have drives firmware forced down to SATA I speeds, it is now showing the correct speed. Having it show a 3 Gbps rate would make the drive faster, just like this product would make your car go faster:
http://www.youtube.com/watch?v=N8eOTGhTpDY
Only when the data resides wholly within the cache. The instant you step outside of the bounds of the cache (and it isn't that big) you will be confined to < SATA1 speeds, regardless of whether the mechanical drive is linked @ a SATA I or SATA II rate.
I have to say i was chocked when i saw the real speed of my harddrive. It is a sata 2 hdd working with a firmware limitation to sata 1. It is the most stupid thing i have ever seen a manufacture do. There's no reason to do that.
As everybody (including me) like numbers... results, i've mentioned before a benchmark, so here it is.
This benchmark was done in my desktop. It is not conclusive and as it's possible to see in the numbers, it is not reliable, it just give us an ideia of the whole thing.
It's important to pay attention in the hdd config in my pc:
2 x samsung 160gb HDD, RAID0 stripped in 64kb block size;
JMB36X RAID controller connected with 1-Lane 2.5Gbps PCI Express bus
So, as it is a pcie controller, the results in the benchmark may be affected by the max transfer rate between the raid controller and the motherboard chipset.
General results:
Sata 1 test 1
Sata 1 test 2
Sata 2 test 1
Sata 2 test 2
Sata 2 test 3
As everybody (including me) like numbers... results, i've mentioned before a benchmark, so here it is.
This benchmark was done in my desktop. It is not conclusive and as it's possible to see in the numbers, it is not reliable, it just give us an ideia of the whole thing.
It's important to pay attention in the hdd config in my pc:
2 x samsung 160gb HDD, RAID0 stripped in 64kb block size;
JMB36X RAID controller connected with 1-Lane 2.5Gbps PCI Express bus
So, as it is a pcie controller, the results in the benchmark may be affected by the max transfer rate between the raid controller and the motherboard chipset.
General results:
Sata 1 test 1
Sata 1 test 2
Sata 2 test 1
Sata 2 test 2
Sata 2 test 3
The results don't back your claim that there's a significant speed difference with the drives in use for SATA I or SATA II.
yeah, i could see that.
but, the best way to get more reliable results, is to benchmarch a sata 2 hdd on the macbook, an then, limit its speed using jumper to sata 1 and compare the results.
wow, i had to discuss so much and just now one human being realised what i was trying to say... Apple shipping slow hdds pricing as gold.....
Agreed. After a major OS X update, my boot speed has always slowed down for the next reboot or two. After that, it's back to 15 seconds with my SSD
HDD benchmarks
Hello. On my Mac after upgrading to 10.6.3 i received the same message in System Profiler (Negotiated Speed Link 1.5 Gigabit). Yesterday I ran some benchmarks on two different HDDs after fresh install of Snow Leopard. I used XBench for testing. I own Macbook Pro 5,1 (15", Late 2008, 2.53 Mhz)
1. Original HDD from Apple - Hitachi HTS543232L9SA02. I ran 3 times the XBench on 10.6.2 so the scores are (39.7, 39.43, 40.29). I ran 3 times the XBench on 10.6.3 so the scores are (38.77, 37.48, 38.57).
2. Second HDD - Fujitsu MHZ2250BH G1. Same 3 tests on 10.6.2 and results (47.02, 48.24, 50.37). Again 3 tests on 10.6.3 - (49.45, 52.49, 47.49).
So my HDD speed has not slowed down after the upgrade. So I assume this only matters for SSDs. My HDDs can never reach 3 Gigabit. I think (not sure) 1.5 Gigabit has less power consumption vs 3 Gigabit mode. So maybe Apple decided to force HDD that cannot reach 3 Gigabit to use 1.5 Gigabit for more power economy. Can somebody proof the power consumption of 1.5 and 3 Gigabit?
Hello. On my Mac after upgrading to 10.6.3 i received the same message in System Profiler (Negotiated Speed Link 1.5 Gigabit). Yesterday I ran some benchmarks on two different HDDs after fresh install of Snow Leopard. I used XBench for testing. I own Macbook Pro 5,1 (15", Late 2008, 2.53 Mhz)
1. Original HDD from Apple - Hitachi HTS543232L9SA02. I ran 3 times the XBench on 10.6.2 so the scores are (39.7, 39.43, 40.29). I ran 3 times the XBench on 10.6.3 so the scores are (38.77, 37.48, 38.57).
2. Second HDD - Fujitsu MHZ2250BH G1. Same 3 tests on 10.6.2 and results (47.02, 48.24, 50.37). Again 3 tests on 10.6.3 - (49.45, 52.49, 47.49).
So my HDD speed has not slowed down after the upgrade. So I assume this only matters for SSDs. My HDDs can never reach 3 Gigabit. I think (not sure) 1.5 Gigabit has less power consumption vs 3 Gigabit mode. So maybe Apple decided to force HDD that cannot reach 3 Gigabit to use 1.5 Gigabit for more power economy.
Can somebody proof the power consumption of 1.5 and 3 Gigabit?So my HDD speed has not slowed down after the upgrade. So I assume this only matters for SSDs. My HDDs can never reach 3 Gigabit. I think (not sure) 1.5 Gigabit has less power consumption vs 3 Gigabit mode. So maybe Apple decided to force HDD that cannot reach 3 Gigabit to use 1.5 Gigabit for more power economy.
It's a nice theory, but not factually accurate. My 250GB WD Scorpio is 5400 RPM and there's no way in hell it can hit close to the 350++ MB/second it would require to saturate the interface. It can't even hit the 175+/second it would take to saturate the SATA I interface.
Yet it is linked @ 3 Gbit/second. See attachments.
Which reminds me that I need to update my signature
Attachments
HDD benchmarks
Actually I benchmarked, because I wanted to check if there is a slowdown when upgrading from 10.6.2 to 10.6.3 - no there isn't. This is what I assume from my benchmark. I don't have SSD to benchmark it, so I tried two different 5400 HDDs on both 10.6.2 and 10.6.3. Have you tried to take out the SSD and to start your Mac only with the WD? Maybe if you have SSD connected to your system, 10.6.3 automatically put all SATA to 3.0 Gbit.
Actually I benchmarked, because I wanted to check if there is a slowdown when upgrading from 10.6.2 to 10.6.3 - no there isn't. This is what I assume from my benchmark. I don't have SSD to benchmark it, so I tried two different 5400 HDDs on both 10.6.2 and 10.6.3. Have you tried to take out the SSD and to start your Mac only with the WD? Maybe if you have SSD connected to your system, 10.6.3 automatically put all SATA to 3.0 Gbit.
It's a nice theory, but not factually accurate. My 250GB WD Scorpio is 5400 RPM and there's no way in hell it can hit close to the 350++ MB/second it would require to saturate the interface. It can't even hit the 175+/second it would take to saturate the SATA I interface.
Yet it is linked @ 3 Gbit/second. See attachments.
Which reminds me that I need to update my signature
I think what is happening here is that there are three different types of drives being used by everyone. SATA I drives, SATA II drives that have been firmware forced down to SATA I by Apple, and then SATA II drives.
SATA I drives and SATA II drives firmware forced to SATA I, will show the link speed as 1.5 Gbps. SATA II drives will show the link speed as 3 Gbps, even though they cannot make use of this throughput rate (it is more of a theoretical limit). As you have mentioned, there is no performance loss on SATA II drives that have been forced down to SATA I. My impression is that Apple firmware forced the SATA II drives to SATA I for compatibility reasons (like on my mid-2009 MBP 13": if my aftermarket SATA II drive had the Apple SATA I firmware, I wouldn't be having any issue right now with EFI 1.7+). They decided forcing SATA II drives down to SATA I would not be an issue because no performance loss would result.
The only difference users are seeing now, is the true link speed actually being operated at, instead of just the maximum link speed offered by the interface.
I also have a WD Scorpio Blue 500GB but in a Macbook Pro 5,1 (15", Late 2008, 2.53 Mhz) so not the stock drive. Mine is showing 3/3 Gigabit.
Attachments
I understand what you're saying OldMike, and that makes sense -- when it comes to mechanical HDD drives. Since they can never reach above SATA I speeds, plus SATA II has reliability problems with the Nvidia chipset (why only on Mac's BTW ??), what harm is keeping it at SATA I. Idealy Apple/Nvidia would just fix the Mac firmware and Apple would not need to disable the SATA II capability of its stock HD's.
However, when it comes to Apple-supplied SSD's, why on earth would they force them to SATA I?? SSD's are the one thing that can use above SATA I speeds. And according to djrod, he has an Apple-supplied SSD (which are only Samsung SSDs last I heard), yet STILL has it forced to SATA I speeds... Why??
However, when it comes to Apple-supplied SSD's, why on earth would they force them to SATA I?? SSD's are the one thing that can use above SATA I speeds. And according to djrod, he has an Apple-supplied SSD (which are only Samsung SSDs last I heard), yet STILL has it forced to SATA I speeds... Why??
OK, after reading this thread, and not having upgraded to 10.6.3 yet, I decided to capture some benchmarks as I did the install.
First off, I am running an OCZ Vertex SSD as my os drive.
on 10.6.2, my XBench disk marks were:
Results 214.87
System Info
Xbench Version 1.3
System Version 10.6.2 (10C540)
Physical RAM 8192 MB
Model MacBookPro5,5
Drive Type OCZ-VERTEX
Disk Test 214.87
Sequential 207.81
Uncached Write 232.22 142.58 MB/sec [4K blocks]
Uncached Write 232.24 131.40 MB/sec [256K blocks]
Uncached Read 119.98 35.11 MB/sec [4K blocks]
Uncached Read 434.51 218.38 MB/sec [256K blocks]
Random 222.42
Uncached Write 75.58 8.00 MB/sec [4K blocks]
Uncached Write 322.87 103.36 MB/sec [256K blocks]
Uncached Read 2316.77 16.42 MB/sec [4K blocks]
Uncached Read 817.64 151.72 MB/sec [256K blocks]
AFTER I upgraded to 10.6.3, My SATA link and negotiated speed was verified at 3 gigabit. The XBench test yielded:
Results 253.50
System Info
Xbench Version 1.3
System Version 10.6.3 (10D573)
Physical RAM 8192 MB
Model MacBookPro5,5
Drive Type OCZ-VERTEX
Disk Test 253.50
Sequential 223.64
Uncached Write 263.56 161.82 MB/sec [4K blocks]
Uncached Write 244.01 138.06 MB/sec [256K blocks]
Uncached Read 131.91 38.61 MB/sec [4K blocks]
Uncached Read 414.50 208.33 MB/sec [256K blocks]
Random 292.55
Uncached Write 101.82 10.78 MB/sec [4K blocks]
Uncached Write 455.36 145.78 MB/sec [256K blocks]
Uncached Read 2212.79 15.68 MB/sec [4K blocks]
Uncached Read 830.71 154.14 MB/sec [256K blocks]
SOOOO... if that minor performance gain is due to 10.6.3 --- then thanks Apple! My guess is if I run the XBench 10 more times, half will be faster thatn 10.6.2 and half will be faster.
First off, I am running an OCZ Vertex SSD as my os drive.
on 10.6.2, my XBench disk marks were:
Results 214.87
System Info
Xbench Version 1.3
System Version 10.6.2 (10C540)
Physical RAM 8192 MB
Model MacBookPro5,5
Drive Type OCZ-VERTEX
Disk Test 214.87
Sequential 207.81
Uncached Write 232.22 142.58 MB/sec [4K blocks]
Uncached Write 232.24 131.40 MB/sec [256K blocks]
Uncached Read 119.98 35.11 MB/sec [4K blocks]
Uncached Read 434.51 218.38 MB/sec [256K blocks]
Random 222.42
Uncached Write 75.58 8.00 MB/sec [4K blocks]
Uncached Write 322.87 103.36 MB/sec [256K blocks]
Uncached Read 2316.77 16.42 MB/sec [4K blocks]
Uncached Read 817.64 151.72 MB/sec [256K blocks]
AFTER I upgraded to 10.6.3, My SATA link and negotiated speed was verified at 3 gigabit. The XBench test yielded:
Results 253.50
System Info
Xbench Version 1.3
System Version 10.6.3 (10D573)
Physical RAM 8192 MB
Model MacBookPro5,5
Drive Type OCZ-VERTEX
Disk Test 253.50
Sequential 223.64
Uncached Write 263.56 161.82 MB/sec [4K blocks]
Uncached Write 244.01 138.06 MB/sec [256K blocks]
Uncached Read 131.91 38.61 MB/sec [4K blocks]
Uncached Read 414.50 208.33 MB/sec [256K blocks]
Random 292.55
Uncached Write 101.82 10.78 MB/sec [4K blocks]
Uncached Write 455.36 145.78 MB/sec [256K blocks]
Uncached Read 2212.79 15.68 MB/sec [4K blocks]
Uncached Read 830.71 154.14 MB/sec [256K blocks]
SOOOO... if that minor performance gain is due to 10.6.3 --- then thanks Apple! My guess is if I run the XBench 10 more times, half will be faster thatn 10.6.2 and half will be faster.
Not all SSDs are faster than SATA I speeds. In fact, I don't think the Apple/Samsung SSDs are fast enough. The Intel, Sandforce, and some others will need SATA II though...
I know my drive isnt 'top of line' but it appears its running at 1.5Gbps as well.
Early-2009 MBP 17" 2.66 320GB
according to fujistu...er toshiba's website it states it may run at 1.5Gbps or 3Gbps.
http://sdd.toshiba.com/techdocs/MHZ2xxxBH.pdf
http://sdd.toshiba.com/main.aspx?Pa...MHZ2xxxBHSeries/MHZ2xxxBHSeriesSpecifications
I plan on throwing in a 640GB WD anyway...so that will probably change.
but interesting to know!
Early-2009 MBP 17" 2.66 320GB
according to fujistu...er toshiba's website it states it may run at 1.5Gbps or 3Gbps.
http://sdd.toshiba.com/techdocs/MHZ2xxxBH.pdf
http://sdd.toshiba.com/main.aspx?Pa...MHZ2xxxBHSeries/MHZ2xxxBHSeriesSpecifications
I plan on throwing in a 640GB WD anyway...so that will probably change.
but interesting to know!