September 30, 2025 • 24 mins
What if the real cause of your random reboots isn’t the GPU at all—but the power plan behind it? We take you end to end through a stability-first build, starting with the underrated hero of every system: clean, properly sized power. You’ll learn how to calculate wattage with 25–30% headroom, navigate 80 Plus efficiency tiers, and safely adopt ATX 3.0 with the 12VHPWR connector—no sharp bends, modular cable sanity, and the UPS/surge stack that prevents nasty surprises when the lights flicker.
From there, we shift into storage strategy that balances speed and safety. HDD, SATA SSD, and NVMe each earn their place, and we break down RAID 0/1/5/6/10 in plain language so you can pick the right array for your workload. We underline a hard truth: RAID protects against disk failure, not human error, so versioned offsite backups remain non-negotiable. Real-world stories—including a painful RAID 5 rebuild gone wrong—highlight why RAID 6 and RAID 10 matter for bigger or busier systems.
Memory and CPU round out the backbone. We simplify DDR4 vs DDR5, explain how frequency and CAS affect real latency, and show why matched pairs and dual channel deliver the performance you paid for. You’ll get quick wins like enabling XMP/EXPO, when ECC is worth it, and how to troubleshoot training hiccups. Then we open the CPU: cores, threads, cache, sockets, chipsets, and why firmware comes before hardware when upgrades fail to post. Cooling decisions—air, AIO, or custom—tie directly to performance ceilings, along with safe overclock/undervolt practices and thermal targets under sustained load.
By the end, you’ll have a practical checklist to build smarter, troubleshoot faster, and feel ready for the CompTIA A+ exam: power headroom, cable stewardship, airflow planning, RAID with backups, memory matching, BIOS compatibility, and validation testing. If this guide helps you ship a rock-solid PC, share it with a friend, leave a quick review, and hit follow so you never miss the next masterclass.
If you want to help me with my research please e-mail me.
Professorjrod@gmail.com
If you want to join my question/answer zoom class e-mail me at
Professorjrod@gmail.com
Art By Sarah/Desmond
Music by Joakim Karud
Little chacha Productions
Juan Rodriguez can be reached at
TikTok @ProfessorJrod
ProfessorJRod@gmail.com
@Prof_JRod
Instagram ProfessorJRod
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
SPEAKER_00 (00:27):
And welcome to Technology Tap.
I'm Professor J.
Rod.
In this episode from WhatsApp toCores, Building a Stable PC.
Let's get into it.
(01:07):
Welcome back to Technology Tap,the show where we bridge theory
and real-world technology.
I'm your host, Professor J-Rod,and today we're building from
the ground up the power,storage, and memory in N CPU to
form the backbone of everystable system.
By the end of this masterclass,you'll understand how to choose
efficient PSUs, configure ratearrays for both performance and
(01:30):
protection, tune memory forbandwidth and reliability, and
select CPUs that keep pace withmodern workloads.
This episode is your roadmap tobuild smarter, troubleshooting
faster, and passing that ComtiaA exam with confidence.
Let's get started.
Every stable build begins withpower.
(01:52):
It's the one component most newtechnicians overlook, yet it
dictates everything performance,longevity, and safety.
Start by adding your system'stotal wattage, CPU, GPU, drives,
fans, then add 25 to 30%headroom.
This overhead ensures your PSUcan handle transit spikes, those
certain surges in current whenGPUs boost or CPUs ramp up under
(02:18):
heavy loads.
If your rig pulls 500 wattsunder stress, aim for at least
650.
It's better to have extracapacity than risk and
stability.
Now check efficiency.
The 80% certification programmeasures how much power converts
it to usable DC power, and it'sinto four different categories.
(02:39):
Bronze is 82 to 85%, budgetfriendly, hotter, noisier.
Silver is 85 to 88%, mid-rangebuilds, gold is 87 to 90%, ideal
for most users, efficient,quiet, cool, and platinum
titanium is 90% plus premiumservers or continuous workloads.
(02:59):
Higher efficiency means lesswaste, heat, lower noise, and
reduced energy costs.
Next, consider ATX 3.0, the newPSU standard.
It introduced the 12 VHPWRconnector, a single 16-pin cable
delivering up to 600 watts fornext generation GPUs like
(03:20):
NVIDIA's RTX 4000 series.
It replaces multiple 8-pinconnectors and better manages
transistance loads.
Warning (03:28):
avoid bending the 12-volt HPWR cable sharply near
the connector.
Keep at least 35mm clearance toprevent overhead.
Choose your cable design wisely.
Fully modular, every cabledetach, best for airflow and
cable management.
Semi-modular, fixed core cablesand detachable peripherals.
(03:51):
Non-modular, all cables fixed,cheaper but cluttered.
Now, protect that power.
Pair your PSUs with a UPS,uninterruptible power source to
guard against blackouts and adda surge protector for voltage
spikes.
A UPS gives you time to savework and shut down safely and
gracefully.
(04:12):
Essential for office and labenvironments.
Keep your system power statestoo.
RAM powered, quick uh resume,hibernate, RAM safe to disk,
zero power to use.
Hibernate sleep combines bothsafe combines both safe tuning
(04:33):
power loss.
Choose sleep for short breaks,hibernate for longer downtime.
Finally, cooling.
Heat is the enemy ofelectronics.
Follow this airflow rule.
Front bottom intakes, rear topexhaust.
Apply thermal pace sparingly, aP-size dot in the center spreads
evenly under pressure.
Don't put it like if you'remaking a sandwich and you're
(04:55):
putting mayonnaise on it.
That's not how it works.
Let's not do that, guys.
Liquid cooling handles heavyloads quietly, but check for
pumps and that if and fittingsregularly because it might leak.
Story time.
A student once paired a 4090 GPUwith a 500-watt bronze PSU.
Random reboots plagued everygaming session.
(05:17):
After upgrading to a 750-wattgold ATX 3.0, stability was
instant.
Remember, power isn't whereyou're where you cut corners,
it's where you invest inreliability.
Storage devices and RAID.
Let's start.
Let's store some data.
Storage determines not onlycapacity but also speed, safety,
(05:39):
and system behavior.
Three main storage type HTD,spinning platters, slower, cost
efficient.
SSD, flashbase, silent, faster,shock resistant.
NVME, PCIE interface, extremelyextreme speed for OS and active
files.
But drives are only half thestory.
(05:59):
How do you organize definesperformance and redundancy?
That's where RAID, reductant,redundant array of independent
disks comes in.
Let's explore each RAID level,its mechanicalisms, advantages,
and disadvantages.
RAID zero, stripping.
How does it work?
Data splits across two or moredrives for parallel read writes.
What does that do?
(06:20):
What's the advantage?
Maximum performance.
Ideal for large sequential readwrites, 100% capacity
utilization.
Essentially, you're using twodrives to act like one.
So that each one takes turnreading and writing.
So it performances it performsreally fast.
Disadvantage.
No redundancy.
(06:41):
If one drive fails, all data islost.
Rebuild impossible.
Backup mandatory.
Best for scratched disk ortemporary rendering spaces for
non-critical data.
RAID 1, mirroring.
How it works.
Data duplicating on two drives.
What's the advantage?
High fault tolerance.
Survive one drive failure.
(07:02):
Easy recovery, just replace thefailed drive and rebuild.
Disadvantage (07:06):
50% capacity efficiency, one drive's worth of
storage, slightly slower rights.
Best for OS drive, smallbusiness servers, mission
critical boot volumes.
The one thing about mirroring isyou're paying for two drives,
but you're only ever going touse one.
So you have four two terabytedrives, you had eight terabytes
(07:30):
of space.
You're only always going to usefour.
That's the bad thing aboutmirroring.
But the good thing is you have abackup.
(07:50):
Survives one drive failure,efficient use of space.
What's the disadvantage?
Slower rights.
Parity is calculation, so it'sslower.
Long rebuilds, increased risk ofa second failure, which you have
another drive that fails, you'regoing to have total loss.
Best for file servers, mediaarchives, environments,
balancing speed, and safety.
(08:12):
Now, in the CompTIA books, inthe in the instructions, in the
data, and all the stuff that youfind from CompTIA, they will
tell you that if your raid dies,wait until it's downtime to
replace it.
In my experience, I replace itright away.
Because you never if you haverate five, because you one drive
(08:35):
already died.
Who knows when the other onemight die?
So you change it right away.
You don't, you know, if it dieson Tuesday, you don't wait till
Saturday to change it.
Right?
You want to change it now.
So, and the literature tells youwait, you don't really want to
wait.
Alright, rate six, strippingplus dual parity.
Here's how it works (08:54):
two parity blocks stored across four or
more drives.
The advantage here is two drivescan fail.
Right?
Safer during rebuild.
The disadvantage is slowerrights due to dual parity,
reduces usable capacity, bestfor mission critical storage,
large arrays, enterprise Nas ornetwork attached storage.
(09:18):
RAID 10, striped mirrors, oneplus zero, high works, mirrors
pairs of drives and then stripsthose pairs.
Advantages it combines the speedof a RAID zero with a redundancy
of RAID one.
Fast rebuilds, only mirroredones are affected.
Excellent performance and faulttolerance.
Disadvantage again, you havefour drives, you're only really
(09:40):
using two, right?
50% capacity, higher cost,double drives, best for
database, virtualization, heavyI.O.
workloads.
The key reminder with RAID isRAIDs protect against hardware
failure, not human error.
Deleting a file, RAID won't saveyou.
Always maintain external orcloud backup.
(10:04):
Storytime.
A design firm trusted RAID 5alone.
One drive failed, then anotherduring the rebuild.
Ten years of work gone.
They switched to RAID 6 plusnightly cloud backup.
Protect data twice, redundancyand backup.
Segment 3, memory mastery.
(10:24):
Memory is your system'sworkplace.
Fast temporary storage foractive data.
The CPU fetches from RAMthousands of times faster than
disk.
RAM generations DDR3 point DDR31.5 volts up to 200 2133 MHz.
(10:45):
That's a legacy system.
We don't usually do DDR3anymore.
DDR4, 1.2 volt up to 3600 MHz,mainstream standard, and DDR5
1.1 volt 4800 to 8400 megahertz.
Do 32-bit sub channels on DIMPMIC.
Each has unique notches.
(11:06):
You cannot put a DDR3 into theDDI5 slot or vice versa.
It would not work because theywon't fit.
Performance depends on frequencyand latency.
Higher megahertz equals morebandwidth.
Lower CLs is a faster firstaxis.
True performance, latencydivided by frequency times 2000.
(11:29):
The lower the number, thebetter.
Channels multiply throughput,single channel 64-bit is
baseline dual 128-bit dualbandwidth and quad 256-bit
servers and workstations.
Use match pairs in color codedslots.
Mismatch modules drop to singlechannel or flex mode reducing
(11:52):
speed.
So you might have four slots andthey're color-coded red, black,
red, black.
You put one in the red, theother one has to go in the red
if you only have two.
If you only have two uh RAMchips.
Now if you have two four and twoeight gig, you put one eight gig
in the red, the other one has togo in the red, and the other two
(12:15):
go into black.
You can't mismatch.
It has to be match pairs.
Exactly match pairs.
ECC versus non-ECC, errorcorrecting code detects and
correct single bits, non-ECC, noerror correction, ECC D support
from CPU and motherboard, idealfor servers, unnecessary from
(12:37):
home PCs.
Never mix ECC and non-ECCtogether.
Now the reason why it's notnecessary from home PC is two
things.
One, they are more expensive.
ECC chips are more expensive.
Two, if your server if yourworkstation crashes because of a
memory issue, just reboot it.
Most of the time that will fixit.
(12:58):
For a server, you have like Xnumber of people connected to
it.
You cannot be rebooting thatserver all the time.
So you want to get the serverECC, error correcting code RAM
chips, so you can correct it onthe fly, and you don't need to
reboot.
You don't need to put these onworkstations.
It doesn't make sense.
You're only wasting money.
(13:21):
Four factors for RAM chips,they're DIM for desktops and so
DIMs for laptops.
RDIM and LR DIM are serversbuffered for stability at high
capacity.
Enable XMP on your Intels or XPOfor AMD profiles and BIOS to
reach rated speedsautomatically.
(13:41):
Unsupported boards may fail postright when they first start up.
Troubleshooting.
Random freezes, reset the RAM.
Blue screen of death, run Mtest86.
Boot issues, try one stick at atime.
One time I bought new RAMsticks, and before you just put
(14:07):
them in and you turn it on, andthe box will automatically
recognize it.
This time it took me like a goodminute of just sitting there
with a black screen on beforethe RAM actually kind of like
connected and the machine turnedon.
But yeah, it was weird.
I thought I thought I got badRAM for a while from Dell.
(14:27):
And then I waited like oneminute and it just turned on.
Just the first time.
It doesn't happen after that.
It was just like I guess itneeded to see to connect with
the bios and update and see youknow notice that the motherboard
that it that it changed, Iguess.
Story time.
A workstation upgrade from an 8gig to 16 gig with mixed brands.
(14:52):
No dual channel, no performancegain.
After installing a match to amatch to 8 gig kit, rendering
speed up 25%.
Matching matters.
So you got to match.
Good.
Memory takeaway capacity equalsmultitaxing, multitasking speed
and latency is responsiveness.
(15:13):
Channels equals bandwidth.
ECC is reliability.
Next, CPU architecture andperformance.
The CPU is your system's brain,executing instructions,
crunching numbers, andcoordinating every process.
Inside every processor is an ALUfor math and logic.
A CU directs operations, aregister, tiny high-speed
(15:36):
storage, and cache L1, 2, and 3lightning fast memory tiers.
Cores are independent processingunit.
Threads are virtual lanescreated by simultaneous
multi-threading or Intel's hyperthreading.
More cores equals bettermultitasking.
Higher gigahertz equals bettersingle thread speed.
(15:56):
The CPU pipeline handles fetch,decode, execute, right back.
Bottom chips uses branchpredictions and out-of-order
execution to keep thosepipelines full.
Instruction sets X86 or X64complex instructions for desktop
ARM risk simplify simplify powerefficient mobile and embedded.
(16:25):
Cash hierarchical L1 is 32 to 64kilobits per core fastest.
L2 256 kilobytes to onemegabytes per core.
L3 shares up to tens ofmegabytes.
More cash reduces trips to RAM,boosting performance, sockets
and compatibility.
(16:46):
Intel has the LAN grid array,pins on the board.
AMD has the pin grid array, pinson or pads on the CPU.
Always match socket, chipsets,and BIOS versions.
Chipsets defines the IOs, PCIElanes, USB ports, overclocking.
(17:06):
Check manufacture, CPU supportlistings before upgrading.
Cooling solutions.
Air coolers, reliable,affordable.
AIO liquid coolers, greatthermals, low noise, custom
loops, maximum cooling, highmaintenance.
Overclocking increases frequencyfor performance but raises heat
(17:28):
and voltage.
Test with Prime 95, ascend benchand keep temps under 85 degrees
Celsius.
Undervolting, lower voltage forquieter, cooler systems.
Common issues, no post, checkEPS, A-Pin, reset CPU, update
BIOS.
Thermal trotting, repaste, fixairflow.
(17:51):
Incompatibility, were wrongsocket or outdated firmware.
A student installed a 13 genIntel CPU in a 600 series board.
Fan spawns, no display, BIOSupdated, fixed it.
Firmware first, hardware second.
Installation recap, liftretention arm.
(18:11):
This is for the CPU, right?
When you're installing a CPU,lift retention arm, and uh align
triangle marker, place CPUgently, lower the lock, apply
P-size paste, mount coolerevent, boot, verify temps, and
core counts.
And the retention of the arm,that's uh that's the that's a
(18:36):
there's something called therewhen you're lifting the arm.
I forgot what that was called.
If anybody remembers, email me.
But there's when you're liftinga retention arm, there's it's
not called retention arm, it'sactually called something else.
I can't think of it.
But if anybody remembers, emailme, professorjrod at gmail.com.
(18:58):
Man, I I can't think of it.
Alright.
CPU takeaways, cores, multitask,cash accelerates, architecture
defines efficiency, and cooliesystem sustain performance.
Compatibility checks preventsheadaches.
Alright, now I'm going to giveyou four Comtias style
questions.
I'm gonna read them and thenread them again, and I'll give
(19:22):
you five seconds to answer them.
Alright, question one (19:25):
which PSU standard introduced the 12 volt
HPWR connector supporting up to600 watts?
A ATX 2.4, B ATX 3.0, C EPS 12volt DSFX.
Here's the question again.
Which PSU standard introduce the12 volt HPWR connector
(19:48):
supporting up to 600 watts?
A ATX 2.4 B ATX 3.0 C EPS 12volt or DSFX?
I'll give you five seconds tothink about it.
543 21.
Alright, the answer is B ATXadds 12 volt HPWR for high
(20:15):
transistance GPUs.
Alright, next question.
Which ray level offers a maximumspeed but zero redundancy?
A ray zero, B RAID 1, C RAID 5,D Ray 6.
Which array level offers amaximum speed but zero
(20:35):
redundancy?
A RAID 0, B RAID 1, C RAID 5, DRAID 6?
I'll give you five seconds toanswer.
5, 4, 3, 2, 1.
And the answer is A ray 0 stripsdata only.
One failure equals total loss.
(20:56):
Alright, next question.
A workstation needs speed andfault tolerance with four
drives.
Which rate is best?
A rate one B rate five, C ratesix, D, rate ten.
I'll read it again.
A workstation needs speed andfault tolerance with four
drives.
Which rate is best?
(21:16):
A rate one, B, rate five, C,rate six, D, rate ten.
Now this is this is uh actuallya good question because what you
need to solve here is twothings.
You see the and is in themiddle.
So you need to solve need speedand fault tolerance.
(21:38):
So whatever answer that youchoose, it has it has to cure
both issues that you're thatit's asking for, which is speed
and fault tolerance.
Now, out of these four, whichone does both speed and fault
tolerance?
And there's only one D Ray 10.
Ray 10 describes stripping andmirroring or performance and
(21:59):
protection.
That's how you have to look atit when they give you an and
question.
You have to look at it.
Well, what are they?
I gotta meet whatever they'reasking before the and and after
the and both conditions must bemet.
Or if there's more, you know, atleast two, right with the word
and.
But yeah.
All right, last question.
After installing a new CPU, thePC powers on but no display.
(22:23):
What's most likely the cause?
A bad RAM, B, insufficient PSU,C outdated BIOS, D faulty GPU.
After installing a new CPU, thePC powers on but no display.
What's the most likely cause?
A bad RAM, B insufficient PSU, Coutdated outdated BIOS or D
(22:45):
faulty GPU.
Now think about it.
I'll give you five seconds.
Five, four, three, two, one.
Alright.
So it's a the clue is new CPU.
So right, so it's not gonna bethe RAM because it's not, and
it's not gonna be a faulty GPU.
So you you you left withinsufficient PSU and outdated
(23:09):
BIOS.
And the answer is C outdatedBIOS.
Newer CPU often require a BIOSupdate to recognize the
microcode.
So all right, well, there youhave it, guys.
I hope you like this chapter onCPU and hard drives,
(23:31):
specifically RAID.
RAID is a big topic on theCompteya exam.
So I hope uh I taught you a lot.
And until next time, keeptapping into technology.
(24:06):
This has been a presentation ofLittle Cha Cha Productions Art
by Sarah, music by Joe Kim.
We are now part of the Pod MatchNetwork.
You can follow me at TikTok atProfessor Jrod at J R O D, or
you can email me at ProfessorJrodj R O D at Gmail.
And welcome to Technology Tap.
I'm Professor J.
Rod.
In this episode from WhatsApp toCores, Building a Stable PC.
Let's get into it.
(01:07):
Welcome back to Technology Tap,the show where we bridge theory
and real-world technology.
I'm your host, Professor J-Rod,and today we're building from
the ground up the power,storage, and memory in N CPU to
form the backbone of everystable system.
By the end of this masterclass,you'll understand how to choose
efficient PSUs, configure ratearrays for both performance and
(01:30):
protection, tune memory forbandwidth and reliability, and
select CPUs that keep pace withmodern workloads.
This episode is your roadmap tobuild smarter, troubleshooting
faster, and passing that ComtiaA exam with confidence.
Let's get started.
Every stable build begins withpower.
(01:52):
It's the one component most newtechnicians overlook, yet it
dictates everything performance,longevity, and safety.
Start by adding your system'stotal wattage, CPU, GPU, drives,
fans, then add 25 to 30%headroom.
This overhead ensures your PSUcan handle transit spikes, those
certain surges in current whenGPUs boost or CPUs ramp up under
(02:18):
heavy loads.
If your rig pulls 500 wattsunder stress, aim for at least
650.
It's better to have extracapacity than risk and
stability.
Now check efficiency.
The 80% certification programmeasures how much power converts
it to usable DC power, and it'sinto four different categories.
(02:39):
Bronze is 82 to 85%, budgetfriendly, hotter, noisier.
Silver is 85 to 88%, mid-rangebuilds, gold is 87 to 90%, ideal
for most users, efficient,quiet, cool, and platinum
titanium is 90% plus premiumservers or continuous workloads.
(02:59):
Higher efficiency means lesswaste, heat, lower noise, and
reduced energy costs.
Next, consider ATX 3.0, the newPSU standard.
It introduced the 12 VHPWRconnector, a single 16-pin cable
delivering up to 600 watts fornext generation GPUs like
(03:20):
NVIDIA's RTX 4000 series.
It replaces multiple 8-pinconnectors and better manages
transistance loads.
Warning (03:28):
avoid bending the 12-volt HPWR cable sharply near
the connector.
Keep at least 35mm clearance toprevent overhead.
Choose your cable design wisely.
Fully modular, every cabledetach, best for airflow and
cable management.
Semi-modular, fixed core cablesand detachable peripherals.
(03:51):
Non-modular, all cables fixed,cheaper but cluttered.
Now, protect that power.
Pair your PSUs with a UPS,uninterruptible power source to
guard against blackouts and adda surge protector for voltage
spikes.
A UPS gives you time to savework and shut down safely and
gracefully.
(04:12):
Essential for office and labenvironments.
Keep your system power statestoo.
RAM powered, quick uh resume,hibernate, RAM safe to disk,
zero power to use.
Hibernate sleep combines bothsafe combines both safe tuning
(04:33):
power loss.
Choose sleep for short breaks,hibernate for longer downtime.
Finally, cooling.
Heat is the enemy ofelectronics.
Follow this airflow rule.
Front bottom intakes, rear topexhaust.
Apply thermal pace sparingly, aP-size dot in the center spreads
evenly under pressure.
Don't put it like if you'remaking a sandwich and you're
(04:55):
putting mayonnaise on it.
That's not how it works.
Let's not do that, guys.
Liquid cooling handles heavyloads quietly, but check for
pumps and that if and fittingsregularly because it might leak.
Story time.
A student once paired a 4090 GPUwith a 500-watt bronze PSU.
Random reboots plagued everygaming session.
(05:17):
After upgrading to a 750-wattgold ATX 3.0, stability was
instant.
Remember, power isn't whereyou're where you cut corners,
it's where you invest inreliability.
Storage devices and RAID.
Let's start.
Let's store some data.
Storage determines not onlycapacity but also speed, safety,
(05:39):
and system behavior.
Three main storage type HTD,spinning platters, slower, cost
efficient.
SSD, flashbase, silent, faster,shock resistant.
NVME, PCIE interface, extremelyextreme speed for OS and active
files.
But drives are only half thestory.
(05:59):
How do you organize definesperformance and redundancy?
That's where RAID, reductant,redundant array of independent
disks comes in.
Let's explore each RAID level,its mechanicalisms, advantages,
and disadvantages.
RAID zero, stripping.
How does it work?
Data splits across two or moredrives for parallel read writes.
What does that do?
(06:20):
What's the advantage?
Maximum performance.
Ideal for large sequential readwrites, 100% capacity
utilization.
Essentially, you're using twodrives to act like one.
So that each one takes turnreading and writing.
So it performances it performsreally fast.
Disadvantage.
No redundancy.
(06:41):
If one drive fails, all data islost.
Rebuild impossible.
Backup mandatory.
Best for scratched disk ortemporary rendering spaces for
non-critical data.
RAID 1, mirroring.
How it works.
Data duplicating on two drives.
What's the advantage?
High fault tolerance.
Survive one drive failure.
(07:02):
Easy recovery, just replace thefailed drive and rebuild.
Disadvantage (07:06):
50% capacity efficiency, one drive's worth of
storage, slightly slower rights.
Best for OS drive, smallbusiness servers, mission
critical boot volumes.
The one thing about mirroring isyou're paying for two drives,
but you're only ever going touse one.
So you have four two terabytedrives, you had eight terabytes
(07:30):
of space.
You're only always going to usefour.
That's the bad thing aboutmirroring.
But the good thing is you have abackup.
(07:50):
Survives one drive failure,efficient use of space.
What's the disadvantage?
Slower rights.
Parity is calculation, so it'sslower.
Long rebuilds, increased risk ofa second failure, which you have
another drive that fails, you'regoing to have total loss.
Best for file servers, mediaarchives, environments,
balancing speed, and safety.
(08:12):
Now, in the CompTIA books, inthe in the instructions, in the
data, and all the stuff that youfind from CompTIA, they will
tell you that if your raid dies,wait until it's downtime to
replace it.
In my experience, I replace itright away.
Because you never if you haverate five, because you one drive
(08:35):
already died.
Who knows when the other onemight die?
So you change it right away.
You don't, you know, if it dieson Tuesday, you don't wait till
Saturday to change it.
Right?
You want to change it now.
So, and the literature tells youwait, you don't really want to
wait.
Alright, rate six, strippingplus dual parity.
Here's how it works (08:54):
two parity blocks stored across four or
more drives.
The advantage here is two drivescan fail.
Right?
Safer during rebuild.
The disadvantage is slowerrights due to dual parity,
reduces usable capacity, bestfor mission critical storage,
large arrays, enterprise Nas ornetwork attached storage.
(09:18):
RAID 10, striped mirrors, oneplus zero, high works, mirrors
pairs of drives and then stripsthose pairs.
Advantages it combines the speedof a RAID zero with a redundancy
of RAID one.
Fast rebuilds, only mirroredones are affected.
Excellent performance and faulttolerance.
Disadvantage again, you havefour drives, you're only really
(09:40):
using two, right?
50% capacity, higher cost,double drives, best for
database, virtualization, heavyI.O.
workloads.
The key reminder with RAID isRAIDs protect against hardware
failure, not human error.
Deleting a file, RAID won't saveyou.
Always maintain external orcloud backup.
(10:04):
Storytime.
A design firm trusted RAID 5alone.
One drive failed, then anotherduring the rebuild.
Ten years of work gone.
They switched to RAID 6 plusnightly cloud backup.
Protect data twice, redundancyand backup.
Segment 3, memory mastery.
(10:24):
Memory is your system'sworkplace.
Fast temporary storage foractive data.
The CPU fetches from RAMthousands of times faster than
disk.
RAM generations DDR3 point DDR31.5 volts up to 200 2133 MHz.
(10:45):
That's a legacy system.
We don't usually do DDR3anymore.
DDR4, 1.2 volt up to 3600 MHz,mainstream standard, and DDR5
1.1 volt 4800 to 8400 megahertz.
Do 32-bit sub channels on DIMPMIC.
Each has unique notches.
(11:06):
You cannot put a DDR3 into theDDI5 slot or vice versa.
It would not work because theywon't fit.
Performance depends on frequencyand latency.
Higher megahertz equals morebandwidth.
Lower CLs is a faster firstaxis.
True performance, latencydivided by frequency times 2000.
(11:29):
The lower the number, thebetter.
Channels multiply throughput,single channel 64-bit is
baseline dual 128-bit dualbandwidth and quad 256-bit
servers and workstations.
Use match pairs in color codedslots.
Mismatch modules drop to singlechannel or flex mode reducing
(11:52):
speed.
So you might have four slots andthey're color-coded red, black,
red, black.
You put one in the red, theother one has to go in the red
if you only have two.
If you only have two uh RAMchips.
Now if you have two four and twoeight gig, you put one eight gig
in the red, the other one has togo in the red, and the other two
(12:15):
go into black.
You can't mismatch.
It has to be match pairs.
Exactly match pairs.
ECC versus non-ECC, errorcorrecting code detects and
correct single bits, non-ECC, noerror correction, ECC D support
from CPU and motherboard, idealfor servers, unnecessary from
(12:37):
home PCs.
Never mix ECC and non-ECCtogether.
Now the reason why it's notnecessary from home PC is two
things.
One, they are more expensive.
ECC chips are more expensive.
Two, if your server if yourworkstation crashes because of a
memory issue, just reboot it.
Most of the time that will fixit.
(12:58):
For a server, you have like Xnumber of people connected to
it.
You cannot be rebooting thatserver all the time.
So you want to get the serverECC, error correcting code RAM
chips, so you can correct it onthe fly, and you don't need to
reboot.
You don't need to put these onworkstations.
It doesn't make sense.
You're only wasting money.
(13:21):
Four factors for RAM chips,they're DIM for desktops and so
DIMs for laptops.
RDIM and LR DIM are serversbuffered for stability at high
capacity.
Enable XMP on your Intels or XPOfor AMD profiles and BIOS to
reach rated speedsautomatically.
(13:41):
Unsupported boards may fail postright when they first start up.
Troubleshooting.
Random freezes, reset the RAM.
Blue screen of death, run Mtest86.
Boot issues, try one stick at atime.
One time I bought new RAMsticks, and before you just put
(14:07):
them in and you turn it on, andthe box will automatically
recognize it.
This time it took me like a goodminute of just sitting there
with a black screen on beforethe RAM actually kind of like
connected and the machine turnedon.
But yeah, it was weird.
I thought I thought I got badRAM for a while from Dell.
(14:27):
And then I waited like oneminute and it just turned on.
Just the first time.
It doesn't happen after that.
It was just like I guess itneeded to see to connect with
the bios and update and see youknow notice that the motherboard
that it that it changed, Iguess.
Story time.
A workstation upgrade from an 8gig to 16 gig with mixed brands.
(14:52):
No dual channel, no performancegain.
After installing a match to amatch to 8 gig kit, rendering
speed up 25%.
Matching matters.
So you got to match.
Good.
Memory takeaway capacity equalsmultitaxing, multitasking speed
and latency is responsiveness.
(15:13):
Channels equals bandwidth.
ECC is reliability.
Next, CPU architecture andperformance.
The CPU is your system's brain,executing instructions,
crunching numbers, andcoordinating every process.
Inside every processor is an ALUfor math and logic.
A CU directs operations, aregister, tiny high-speed
(15:36):
storage, and cache L1, 2, and 3lightning fast memory tiers.
Cores are independent processingunit.
Threads are virtual lanescreated by simultaneous
multi-threading or Intel's hyperthreading.
More cores equals bettermultitasking.
Higher gigahertz equals bettersingle thread speed.
(15:56):
The CPU pipeline handles fetch,decode, execute, right back.
Bottom chips uses branchpredictions and out-of-order
execution to keep thosepipelines full.
Instruction sets X86 or X64complex instructions for desktop
ARM risk simplify simplify powerefficient mobile and embedded.
(16:25):
Cash hierarchical L1 is 32 to 64kilobits per core fastest.
L2 256 kilobytes to onemegabytes per core.
L3 shares up to tens ofmegabytes.
More cash reduces trips to RAM,boosting performance, sockets
and compatibility.
(16:46):
Intel has the LAN grid array,pins on the board.
AMD has the pin grid array, pinson or pads on the CPU.
Always match socket, chipsets,and BIOS versions.
Chipsets defines the IOs, PCIElanes, USB ports, overclocking.
(17:06):
Check manufacture, CPU supportlistings before upgrading.
Cooling solutions.
Air coolers, reliable,affordable.
AIO liquid coolers, greatthermals, low noise, custom
loops, maximum cooling, highmaintenance.
Overclocking increases frequencyfor performance but raises heat
(17:28):
and voltage.
Test with Prime 95, ascend benchand keep temps under 85 degrees
Celsius.
Undervolting, lower voltage forquieter, cooler systems.
Common issues, no post, checkEPS, A-Pin, reset CPU, update
BIOS.
Thermal trotting, repaste, fixairflow.
(17:51):
Incompatibility, were wrongsocket or outdated firmware.
A student installed a 13 genIntel CPU in a 600 series board.
Fan spawns, no display, BIOSupdated, fixed it.
Firmware first, hardware second.
Installation recap, liftretention arm.
(18:11):
This is for the CPU, right?
When you're installing a CPU,lift retention arm, and uh align
triangle marker, place CPUgently, lower the lock, apply
P-size paste, mount coolerevent, boot, verify temps, and
core counts.
And the retention of the arm,that's uh that's the that's a
(18:36):
there's something called therewhen you're lifting the arm.
I forgot what that was called.
If anybody remembers, email me.
But there's when you're liftinga retention arm, there's it's
not called retention arm, it'sactually called something else.
I can't think of it.
But if anybody remembers, emailme, professorjrod at gmail.com.
(18:58):
Man, I I can't think of it.
Alright.
CPU takeaways, cores, multitask,cash accelerates, architecture
defines efficiency, and cooliesystem sustain performance.
Compatibility checks preventsheadaches.
Alright, now I'm going to giveyou four Comtias style
questions.
I'm gonna read them and thenread them again, and I'll give
(19:22):
you five seconds to answer them.
Alright, question one (19:25):
which PSU standard introduced the 12 volt
HPWR connector supporting up to600 watts?
A ATX 2.4, B ATX 3.0, C EPS 12volt DSFX.
Here's the question again.
Which PSU standard introduce the12 volt HPWR connector
(19:48):
supporting up to 600 watts?
A ATX 2.4 B ATX 3.0 C EPS 12volt or DSFX?
I'll give you five seconds tothink about it.
543 21.
Alright, the answer is B ATXadds 12 volt HPWR for high
(20:15):
transistance GPUs.
Alright, next question.
Which ray level offers a maximumspeed but zero redundancy?
A ray zero, B RAID 1, C RAID 5,D Ray 6.
Which array level offers amaximum speed but zero
(20:35):
redundancy?
A RAID 0, B RAID 1, C RAID 5, DRAID 6?
I'll give you five seconds toanswer.
5, 4, 3, 2, 1.
And the answer is A ray 0 stripsdata only.
One failure equals total loss.
(20:56):
Alright, next question.
A workstation needs speed andfault tolerance with four
drives.
Which rate is best?
A rate one B rate five, C ratesix, D, rate ten.
I'll read it again.
A workstation needs speed andfault tolerance with four
drives.
Which rate is best?
(21:16):
A rate one, B, rate five, C,rate six, D, rate ten.
Now this is this is uh actuallya good question because what you
need to solve here is twothings.
You see the and is in themiddle.
So you need to solve need speedand fault tolerance.
(21:38):
So whatever answer that youchoose, it has it has to cure
both issues that you're thatit's asking for, which is speed
and fault tolerance.
Now, out of these four, whichone does both speed and fault
tolerance?
And there's only one D Ray 10.
Ray 10 describes stripping andmirroring or performance and
(21:59):
protection.
That's how you have to look atit when they give you an and
question.
You have to look at it.
Well, what are they?
I gotta meet whatever they'reasking before the and and after
the and both conditions must bemet.
Or if there's more, you know, atleast two, right with the word
and.
But yeah.
All right, last question.
After installing a new CPU, thePC powers on but no display.
(22:23):
What's most likely the cause?
A bad RAM, B, insufficient PSU,C outdated BIOS, D faulty GPU.
After installing a new CPU, thePC powers on but no display.
What's the most likely cause?
A bad RAM, B insufficient PSU, Coutdated outdated BIOS or D
(22:45):
faulty GPU.
Now think about it.
I'll give you five seconds.
Five, four, three, two, one.
Alright.
So it's a the clue is new CPU.
So right, so it's not gonna bethe RAM because it's not, and
it's not gonna be a faulty GPU.
So you you you left withinsufficient PSU and outdated
(23:09):
BIOS.
And the answer is C outdatedBIOS.
Newer CPU often require a BIOSupdate to recognize the
microcode.
So all right, well, there youhave it, guys.
I hope you like this chapter onCPU and hard drives,
(23:31):
specifically RAID.
RAID is a big topic on theCompteya exam.
So I hope uh I taught you a lot.
And until next time, keeptapping into technology.
(24:06):
This has been a presentation ofLittle Cha Cha Productions Art
by Sarah, music by Joe Kim.
We are now part of the Pod MatchNetwork.
You can follow me at TikTok atProfessor Jrod at J R O D, or
you can email me at ProfessorJrodj R O D at Gmail.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
Cardiac Cowboys
The heart was always off-limits to surgeons. Cutting into it spelled instant death for the patient. That is, until a ragtag group of doctors scattered across the Midwest and Texas decided to throw out the rule book. Working in makeshift laboratories and home garages, using medical devices made from scavenged machine parts and beer tubes, these men and women invented the field of open heart surgery. Odds are, someone you know is alive because of them. So why has history left them behind? Presented by Chris Pine, CARDIAC COWBOYS tells the gripping true story behind the birth of heart surgery, and the young, Greatest Generation doctors who made it happen. For years, they competed and feuded, racing to be the first, the best, and the most prolific. Some appeared on the cover of Time Magazine, operated on kings and advised presidents. Others ended up disgraced, penniless, and convicted of felonies. Together, they ignited a revolution in medicine, and changed the world.