The modern identification of Banc-y-Warren with Crug Mawr, and the battle site, chiefly rests on this quotation by Lewis*:

Soon after the death Henry I, a memorable battle was fought near Crûg Mawr, a conical hill in this parish [Llangoedmore], between the Welsh, commanded by Grufydd ab Rhys and the English, in which the latter sustained a signal defeat.

Which would lead one to expect that one of the Welsh or Norman histories would include a clear description of the hill and its location. However, this seems not to be the case. History-centred sources tend to identify the battle as being at Aberteif(/v)i. Here is the entry from the B and C versions of the Annales Cambriae* (See Other Works) with a rough translation. The latter parts of B and C may date from 1286 and 1288 respectively:

B: Annus MCXXXVI: Richardus filius Gilberti a Morgano filio Owyni occisus est. Owinus et Catwaladrus filii Grifini exercitum in Karedigeaun movent, et castello Walteri de Bek; et castello Aberystuit, et castello Ricardi de la Mar, et Dineirth destructis, Kairwedros quoque combusserunt, et sic ad propria reversi sunt. Owinus ct Catwaladrus iterum ad Keredigean venerunt, quibus in adjutorium Grifinus filius Resi, et Resus filius Hoeli, et Madocus filius Idnerth, et filii Hoeli ad Abertewy (B: Abertewy; C: Aberteyui} potenter venerunt; quibus ex alia parte resisterunt Stephanus constabularius et filii Geraldi et omnes Franci ab hoste Sabrinae usque ad Meneviam, et Flandrenses de Ros: et proelio coram castellum inito, Franci et Flandrenses in fugam versi capti sunt, occisi sunt, combusti et equorum pedibus conculcati et in fluvio Tewy submersi sunt, plurimi in captivitatem miserrime ducti Walenses vero his ad votum peractis, in sua redierunt, castello Francis remanente
[Year 1136: Richard son of Gilbert killed by Morgan son of Owyn. Owain and Cadwalader the sons of Grifin move in exercises in Ceridigion, and the Castle of Walter de Bek, and the castle of Aberystwyth, and the castle of Richard de la Mar, and Dineirth are destroyed. Caerwedros likewise they have burnt down, and so they return. Owain and Catwalader come a second time to Ceridigion, with the support of Grifin son of Rhys, and Rhys son of Hoel, and Madoc son of Idnerth, and the sons of Hoel, they have powerfully come to Aber Tewy; where from the other part [i.e. against them] resisted Stephen the constable and the sons of Gerald and all the French enemy from the Severn to Menevia [St.David's], and the Flemings of Rhos: and entered into battle infront of the castle, the French and Flemings are captured in flight, killed, burnt, and trampled under the horses feet, and in the river Tewy they are submerged, many are led in miserable captivity by the Welsh to a truely dedicated end[?], to their own they [some?] returned, the castle of the French remained.]

C: MCXXXVI: Grifut filius Res congregato exercitu magno cum omnibus ducibus totius Cambriae pugnavit apud Aberteyn contra Francos et Flandrenses, et victor fuit. Franci autem partim occisi partim in amne submersi fugam fecerunt. Eynaun filius Owein ibi occiditur.
[C: 1136: Grifut son of Rhys convenes a large army with all the leaders of Wales. It has fought at Aberteyn against the French and Flemings, and it has been victorious. Of the French, on the other hand, some are killed, some are submerged by the river - they have run away] Manuscript C footnotes this "Blake Saterndey" i.e. Black Saturday*

Here then, the battle appears to be in front of the castellum (either castle or fortified town, but given the use to describe the property of Richard de la Mare, probably the former).

Here's a translation of the description from the Brut Y Tywysogion by Williams ap Ithel* which is useful in as much as it crams in almost all the manuscript versions of the Brut*. Clarifications are given in brackets { } from Jones* based on the three clear versions of the Brut (Peniarth 20 variants; Red Book of Hergest variants, and the Brenhinedd y Saesson variants) and additions in square-brackets [ ] from Powel's version of the Brut*, which is of various unknown origins**:

1135. The ensuing year, Rickert, son of Gilbert {Richard Fitz-Gilbert}, was slain by Morgan, son of Owain. After that, Owain and Cadwalader, the sons of Gruffudd, son of Cynan, led a large and cruel army into Ceredigion; —the men who were the ornament of all the Britons, their safety, their liberty, and their strength ; the men who were two noble and two generous kings ; {/like two kings, like two generous ones,} two dauntless ones ; two brave lions{/ones}; two blessed ones; two eloquent{/pleasant} ones ; two wise ones; protectors of the churches, and their champions; the defenders of the poor ; the slayers of the foes; the pacifiers of the quarrelsome ; the tamers of antagonists ; the safest refuge to all who should flee to them; the men who were pre-eminent in energies of souls and bodies ; and jointly upholding in unity the whole kingdom of the Britons {/held supremacy over all Wales}. They on the first onset burned the castle of Walter de Bec; and then, having moved their wings, they fought against the castle of Aberystwyth and burned it; and along with Howel, son of Maredudd, and Madog, son of Idnerth, and the two sons of Howel, to wit, Maredudd and Rhys, [and Rhys ap Madawc ap Ednerth] they burned the castle of Eickert de la Mere {Richard de la Mare}, and the castle of Dinerth, and the castle of Caerwedros; and afterwards they returned home. In the close of that year they came a second time into Ceredigion, having with them a numerous army of choice combatants, about six thousand fine infantry, and two thousand one hundred {/2000} {mailed} cavalry in armour. And to their aid came Gruffudd, son of Rhys, and Howel, son of Maredudd of Brecheiniog, and Madog, son of Idnerth, and the two sons of Howel, son of Maredudd [they over-ran the county, as far as Aberteifi, restoring all the former inhabitants to their proper inheritances, and discarding all such strangers as the late Earl of Strygil had placed in the country]. And all those conjointly drew up their troops at Aber Dyvi [See below]. And to oppose them came Stephen the constable of the town, and Robert, son of Martin {Robert Fitz-Martin}, and the sons of Gerald the steward, and William son of Orc {/Odo}[/John], and all the Flemings, and all the marchers, and all the French from Aber Nedd {the Neath} unto Aber Dyvi [See below] [and meeting with the Welch berwixt Aber Ned and Aber Dyfi]. And after joining battle, with cruel fighting on every side, the Flemings and the Normans took to flight, according to their usual custom [trusting too much to the strength of their towns and fortifications, began to look how to save themselves that way]. [But the Welch pressed upon them so hard, that they killed above 3000 men] And after some of them had been killed, and others burned, and the limbs of the horses of others broken {/and others trampled under horses' feet}, and others taken captive, and the greater part, like fools, drowned in the river, and after losing about three thousand of their men, they returned exceedingly sorrowful to their country. After that, Owain and Cadwalader [over-ran the whole county, forcing all the Normans and Flemings to depart the country with all speed, and placing in their room those miserable Welch, who had been so long deprived and kept from their own extates; and after they had weeded the country of those insatiable caterpillars, they] returned, happy and rejoicing, to their country, having obtained the victory honourably with an immense number of prisoners, and spoils, and costly garments and arms. {Pen.20 has a marginalia: "y Sadwrn Du" [Black Saturday]: matching that for the Annales}

Here the situation is even worse, as it isn't even clear the location is Aber Teivi. The various versions of the Brut have different readings*. For the location the troops converge on, Peniarth 20 has Aberteuii, while the Red Book of Hergest version has Aber Teiui, and the Brenhinedd y Saesson has Aberteiui*. While these seem conclusive, the later location between which the Norman forces are collected between runs from Peniarth 20 and the Red Book's Aber Dyvi to the Brenhinedd y Saesson's manuscript B* (Black Book of Basingwerk) Aber Tevi. Moreover, the Brenhinedd y Saesson's manuscripts both have this as where there soliders come to, not from* (somewhat in line with Powel's version), that is there is a line which could argue for Aber Dyvi [Aberdovey] as the location. Note also the difference with the Annales which have "between the Severn and St.David's".

However, it is generally accepted that these Bruts are all based on a Latin original which was, in turn, based on the Annales Cambriae manuscript B. Given this, we are somewhat justified in accepting the location as around Cardigan, and the variations as transcription and translation errors.

The battle also appears in the continuation of the Chronicle of Florence of Worcester by John of Worcester (though the actual addition was by the G² continuator of McGurk* at some time after 1146*). Here's the translation by Thomas Forester* (pretty much matches that of McGurk*):

Another bloody battle was afterwards fought at Cardigan, in the second week of the month of October, in this same year [1136], in which the slaughter was so great that, without reckoning the men who were carried off into captivity, there remained ten thousand women, whose husbands, with numberless children, were either drowned, or burnt, or put to the sword. When the bridge over the Tivy was broken down it was a wretched spectacle to see crowds passing to and fro across a bridge formed by the horrible mass of human corpses and horses drowned in the river.

Here then, there seems some confirmation that the results of the battle at least spilled into the town. Interestingly, this description is very different from that of the Welsh sources, and doesn't seem related to them. As is could have been written within ten years of the event, it is a strong source. Different again is the description in the near-contemporary Gesta Stephani ("The Deeds of [King] Stephen"; c.1148). Here's the translation by Potter*:

When this was spread abroad, namely, that the principal man set over the Welsh [Richard Fitz Gilbert] had fallen; gathering from the different districts into a huge army they invaded his land, dividing themselves into three terrible bands methodically and with a view to war, and surrounded on three sides and routed Richard's own knights with the addition of some others who to the number of three thousand, including foot-soldiers, had assembled to help them from the neighbouring towns and castles; pursuing them in energetic and spirited fashion with shouts and arrows they pitifully slaughtered some, others they massacred by driving them violently into a river, a good number they put in houses and churches to which they set fire and burnt them, and roving as plunderers through the whole district, which extends for thirty-six miles, they left nothing remaining in it; old men they exposed to slaughter or mockery; the young of both sexes they delivered over to chains and captivity; women of any age they shamelessly abandoned to public violation...when the Welsh were troubling the land in this fashion, it seemed to the king that he was striving in vain, in vain pouring out his vast treasure to reduce them to peace; and so, advised by more judicious counsel, he preferred to endure their insolent rebellion for a time, in order that, with fighting at a stand-still and disagreement setting them all at variance, they might either suffer a famine or turn on each other and be exterminated by mutual slaughter.

One might imagine he did.

Anyhow, it can be seen that the Histories are not especially clear about the location of the battle. Other sources are little better. The battle probably features in at least two poems of Y Gogynfeirdd (Info). Though there were battles in the area before this, the people praised in the full versions of the poems fought at Crug Mawr (Cardigan Castle was developed by the Normans in 1093, also 1110*; with early battles there in 1115* [possibly]; 1116* [1113 in MS]; 1145*; 1165* [fought by Rhys, who rebuilt the castle in 1171*]). Here are the two poems in translations from Clancy*:

At Aberteifi, fierce in battle Owain
Blest prince of Britain, land's rightful lord:
Brave men evade him, on hearing him there,
Before Iago's great-grandson, massacre maker

Exultation; Gwalchmai ap Meilyr (Info)

At Aberteifi fallen spears shattered
As at Baddon, clamorous onslaught:
I saw savage troops and stiff red corpses,
It was left to the wolves, their burial;
I saw them abandoned, defenceless
Beneath birds' feet, strong men slain;
I saw their ruin, three hundred dead;
I saw, battle over, bowels on thorns;
I saw dreadful tumult in turmoil,
Troops contending, a rout collapsing.

In Praise of Owen Gwynedd (Info) by Cynddelw Brydydd Mawr (Info)

The poems seem to have been written for an audience that would have a critical understanding of the battle: for example, they give a more realistic total for the dead than the Bruts (though this may be just on the Welsh side), but they still do not locate the battle on the mountain. In defense of a mountain location, however, the latter poem does compare the battle with the battle of Baddon, which is given as a mountainside battle [Monte Badonis] in the section of the Historia on Arthur's battles (See Contents).

The only clear reference that infact connects the battle with Crug Mawr is from the travel diary of Gerald of Wales*. However this is also likely to be a reliable source as, firstly, the site was probably pointed out by a local, and, more importantly, Gerald wrote just within living memory of the battle (51 years later) so the local may have remembered it:

We proceeded on our journey from Cilgerran towards Pont-Stephen [Lampeter], leaving Cruc Mawr, i.e. the great hill, near Aberteivi [Cardigan], on our left hand. On this spot Gruffydd, son of Rhys ap Tewdwr, soon after the death of king Henry I., by a furious onset gained a signal victory against the English army, which, by the murder of the illustrious Richard de Clare, near Abergevenny (before related), had lost its leader and chief. A tumulus is to be seen on the summit of the aforesaid hill, and the inhabitants affirm that it will adapt itself to persons of all stature and that if any armour is left there entire in the evening, it will be found, according to vulgar tradition, broken to pieces in the morning.

That a battle would take place on the mountain rather than next to the castle may seem unlikely, however mountainside battles appear common in both the Bruts and Annales. While there is a plain strategic advantage to the army higher up the mountain (see the four chapters on mountain fighting in von Clausewitz*), it is not clear why forces would leave a fortified area to engage in a battle when this advantage is clearly against them. However, it appears from the sources that armies on both sides at that time encamped on hillsides for defence [see 1159* for a Welsh example; 1165* for a Norman one], and it is possible that an attacking army would aim to surprise an encamped army at rest, either directly or from the lee of the hill. It is also clear that the Welsh were using cavalry in the battle. While cavalry on a hill is apparently a fearful thing to see charging towards you*, they would be extremely unwieldy, and a commander with some guts may have tried to use this to their advantage. The final possibility is that Rhys ap Gruffudd was laying seige to the castle when reinforcements arrived, forcing him to retreat to a more strategic location where his back wasn't against the river. He showed exactly this behaviour in 1159 when he left the castle of Carmarthen, which he had in seige, and moved to the mountain Cefn Rhestr Main when reinforcements arrived to relieve the castle. From this location he managed to force a truce without fighting*, however, he may have had the sucessful battle at Crug Mawr in mind when setting out his location. The move from castle site to mountain would certainly explain the different locations in the Annales and Gerald's account.

(Finally, as an aside, note that the Bruts give the date as 1135, however, they also give the battle as occurring the year after the death of Henry I (indeed, encouraged by it) in 1134. Henry infact died in 1135, making the Annales Cambriae's date for the battle of 1136 correct. The fact that the Annales C and Peniarth 20 label it "Black Saturday", and the Chronicle of John of Worcester gives the date as the second week of October suggested to Lloyd a date of 10th October*)