"Thomas Eg Jørgensen" <thomas@hest.notaplan.com> wrote in message
news:07488D63-9C27-4CD8-905F-D3C2DFD63216@news.elevatesoft.com...
> "Frans van Daalen" <Account@is.invalid> skrev i en meddelelse
> news:F04BBDBB-06E5-4F2A-94AE-7D1A166D174F@news.elevatesoft.com...
>>> << MTRON SSD 32GB: Wile E. Coyote or Road Runner?  >>
>>>
>>> Wow, very impressive.  Especially since 32GB is well within the range of
>>> most smaller databases.
>>>
>> Yes, it's on my list
>>
>> Shame that they did not have a db test.
>>
>
> I'm going to order one of these(maybe the 2,5" depending on deliverytime)
> http://www.king-cart.com/cgi-bin/cart.cgi?store=dvnation&product_name=16GB+3.5+SATA+100MB/s+SSD&exact_match=exact
>
> The combination of 0,1ms seektime and read/write speeds of
> 80MByte/100MByte per second seems quite nice
>
> ...the main purpose is to do some DB testing(BTrieve, MSSQL and DBISAM
> v3)...
>
> I will report back when the disk arrive...
>
> /*Thomas

If you have some real money to spend maybe this one is even better

"
So how fast is the ioDrive?  Flynn said the card has 160 parallel pipelines
that can read data at 800 megabytes per second and write at 600 MB/sec.  He
even proved it by running a Linux drive I/O benchmark.  But for large
corporations running busy databases, operations per second is a much more
important number than bandwidth.
Flynn set the benchmark for the worst case scenario by using small 4K blocks
and then streaming eight simultaneous 1 GB reads and writes.  In that test,
the ioDrive clocked in at 100,000 operations per second.  "That would have
just thrashed a regular hard drive," said Flynn.
"

I/O Benchmark screenshot
http://www.tgdaily.com/index.php?option=com_content&task=blogsection&id=18&Itemid=41&slideshow=20070926&currentPic=3

Frans,

<< If you have some real money to spend maybe this one is even better >>

Damn, that's impressive.  With that type of performance, I can start taking
the weekends off and just say "buy one of these".

--
Tim Young
Elevate Software
www.elevatesoft.com

Tim Young [Elevate Software] wrote:
> Frans,
>
> << If you have some real money to spend maybe this one is even better >>
>
> Damn, that's impressive.  With that type of performance, I can start taking
> the weekends off and just say "buy one of these".
>

Tim,
   You mean you're not going to be giving them away with every EDB
purchase?? Gosh, and to think I had you on my XMAS card list.

Dave

matthew@matthewdelme-jones.delme.com (Matthew Jones) wrote:

> Okay, sometimes good things happen when your mind can have time to think.
> I haven't managed to get around to buying an SSD drive to experiment with
> yet. But I had on my todo list getting a RAM drive to experiment with
> first. Then in a discussion about this, someone said "pity we can't hold
> the whole database in memory", and I went "D'oh!".


For large data bases you'll run out of memory very quickly if you use
only 1 machine. That's why Tim has to get started working on a Cluster
database (like MySQL) where the database can reside in the memory of 32
computers, each with 16gb or more of memory. If any data node goes down,
the others will recover automatically. As I understand it, Tim will have
the weekends off now so he'll have plenty of time to pursue it.<vbg>

Dave

Dave,

<< For large data bases you'll run out of memory very quickly if you use
only 1 machine. That's why Tim has to get started working on a Cluster
database (like MySQL) where the database can reside in the memory of 32
computers, each with 16gb or more of memory. If any data node goes down, the
others will recover automatically. As I understand it, Tim will have the
weekends off now so he'll have plenty of time to pursue it.<vbg> >>

I wish it were that easy.  The difference between MySQL and DBISAM/ElevateDB
can be summed up best in this link:

http://www.mysql.com/company/investors.html

As soon as we get one of those pages, we'll be on top of making ElevateDB
work in a cluster.

--
Tim Young
Elevate Software
www.elevatesoft.com

Tim Young [Elevate Software] wrote:

> Dave,
>
> << For large data bases you'll run out of memory very quickly if you use
> only 1 machine. That's why Tim has to get started working on a Cluster
> database (like MySQL) where the database can reside in the memory of 32
> computers, each with 16gb or more of memory. If any data node goes down, the
> others will recover automatically. As I understand it, Tim will have the
> weekends off now so he'll have plenty of time to pursue it.<vbg> >>
>
> I wish it were that easy.  The difference between MySQL and DBISAM/ElevateDB
> can be summed up best in this link:
>
> http://www.mysql.com/company/investors.html
>
> As soon as we get one of those pages, we'll be on top of making ElevateDB
> work in a cluster.
>

Is that all it takes? I've already copied their HTML page and I'm busy
changing all references of "MySQL" to "ElevateDb". You'll have the new
HTML page for your site in a couple of hours.

Dave

Dave


Swine! I was going to offer to master a page for Tim, I was even going to use some different VC names to totally protect the guilty

Roy Lambert

"Thomas Eg Jørgensen" <thomas@hest.notaplan.com> skrev i en meddelelse
news:07488D63-9C27-4CD8-905F-D3C2DFD63216@news.elevatesoft.com...
> I will report back when the disk arrive...
>

So...i got a visit from the UPS-guy today...nice guy by the way, he
brought me a packageI must remember to send him something for
christmas

Anyways, i've timed some of our "heaviest" software, first on normal
SCSI drives and then on the SSD...but...no significant difference...so
no bottleneck in the disks with our DBISAM software...

Anyone wanna try anything? I have a freshly installed Dell dimension 520
with 3Ghz Pentium4D with 2Gbyte of memory and then finally 16Gbyte SSD
disk....with Vista installed...

/*Thomas

Thomas,

<< Anyways, i've timed some of our "heaviest" software, first on normal SCSI
drives and then on the SSD...but...no significant difference...so no
bottleneck in the disks with our DBISAM software...  >>

If you try to insert a large number of rows into a table with a lot of
indexes, and perform a StartTransaction/Commit every 10,000 rows, you should
be able to start hitting the disk fairly hard, especially if the input is
fairly random.

--
Tim Young
Elevate Software
www.elevatesoft.com

"Thomas Eg Jørgensen" <thomas@hest.notaplan.com> wrote in message
news:B248976E-BAC8-4BB2-9A7F-1B08739F0C8D@news.elevatesoft.com...
> "Thomas Eg Jørgensen" <thomas@hest.notaplan.com> skrev i en meddelelse
> news:07488D63-9C27-4CD8-905F-D3C2DFD63216@news.elevatesoft.com...
>> I will report back when the disk arrive...
>>
>
> So...i got a visit from the UPS-guy today...nice guy by the way, he
> brought me a packageI must remember to send him something for
> christmas
>
> Anyways, i've timed some of our "heaviest" software, first on normal SCSI
> drives and then on the SSD...but...no significant difference...so no
> bottleneck in the disks with our DBISAM software...
>
> Anyone wanna try anything? I have a freshly installed Dell dimension 520
> with 3Ghz Pentium4D with 2Gbyte of memory and then finally 16Gbyte SSD
> disk....with Vista installed...
>
> /*Thomas
What is your definition on "normal" scsi? what disk which interface?